Datasets:

ccdv

/

arxiv-summarization

like 63

in 1978 fisher and de gennes @xcite realized that when thermal fluctuations are restricted by a container a force acts on the walls of the container . since this effect is rather similar to the casimir effect , where the restriction of quantum fluctuations induces a force , it is called `` thermodynamic '' casimir effect . since thermal fluctuations only extend to large scales in the neighbourhood of a continuous phase transitions it is also called `` critical '' casimir effect . recently this effect has attracted much attention , since it could be verified for various experimental systems and quantitative predictions could be obtained from monte carlo simulations of spin models @xcite . in the thermodynamic limit of the three dimensional system , the correlation length , which measures the spatial extension of fluctuations , diverges following the power law @xmath16 where @xmath17 is the reduced temperature and @xmath18 the critical temperature . @xmath19 and @xmath20 are the amplitudes of the correlation length in the high and low temperature phase , respectively . while @xmath19 and @xmath20 depend on the microscopic details of the system , the critical exponent @xmath6 and the ratio @xmath21 are universal . at the critical point also other quantities like the specific heat show a singular behaviour : @xmath22 in the case of the xy universality class that we consider here , the exponent @xmath23 @xcite of the specific heat is negative . therefore the analytic background @xmath24 has to be taken into account . note that the critical exponents of the correlation length and the specific heat are related by the hyperscaling relation @xmath25 , where @xmath26 is the dimension of the system . for reviews on critical phenomena and its modern theory , the renormalization group , see e.g. @xcite . the singular behaviour at the critical point originates from the fact that thermal fluctuations range over all length scales . therefore the behaviour in the neighbourhood of the critical point is modified if the system is confined by a container . a priori thermodynamic quantities are functions of the reduced temperature and the size @xmath4 of the container , assuming a fixed geometry . however the theory of finite size scaling predicts that the physics of the system is governed by the ratio @xmath27 as long as @xmath28 , where @xmath29 is the microscopic scale of the system . in particular if a quantity in the thermodynamic limit behaves as @xmath30 , finite size scaling predicts that @xmath31 , where @xmath32 is the critical exponent of @xmath33 and @xmath34 the correlation length of the bulk system . we can rewrite this equation as @xmath35^{1/\nu})\ ] ] by using ( [ xipower ] ) , which is the form used in the following . note that the function @xmath36 depends on the details of the container . for example for a cube it is different from a thin film . it also depends on the type of boundary conditions that is imposed by the walls of the container on the order parameter of the system . the predictions of finite size scaling theory have been tested in experiments and theoretical studies for various universality classes and confining geometries ; for reviews see @xcite . here we shall focus on thin films in the three dimensional xy universality class , which is shared by the @xmath15-transition of @xmath10he . very precise experimental results for critical exponents and universal amplitude ratios were obtained for this phase transition @xcite . also a large number of experiments on thin films of @xmath10he and @xmath37he-@xmath10he mixtures were performed to probe finite size scaling @xcite . in particular the specific heat @xmath38 of thin films has been studied . the excess specific heat should behave as @xmath39^{1/\nu } ) \;.\ ] ] the reason to study the excess specific heat rather than just the specific heat @xmath40 is to cancel the analytic background @xmath24 . note that the scaling function @xmath41 of the excess specific heat is , up to a constant factor , the second derivative @xmath11 of the scaling function @xmath2 of the excess free energy per area @xmath42 where @xmath43 is the free energy per area of the thin film , @xmath44 the free energy density of the bulk system , @xmath45 the dimension of the system and @xmath5^{1/\nu}$ ] . note that in the case of thin films we consider , following the literature on the thermodynamic casimir effect , free energies per area . we hope that this does not lead to confusion , since in the case of the specific heat , energies per volume are considered . from a thermodynamic point of view , the casimir force per unit area is given by @xmath46 where @xmath4 is the thickness of the film . inserting the finite size scaling ansatz ( [ defineh ] ) for the excess free energy into ( [ definef ] ) we get @xmath47^{1/\nu})]\right ] } { \partial l_0 } \nonumber \\ & = & - k_b t l_0^{-3 } \left[-2 h(t [ l_0/\xi_0]^{1/\nu } ) + \frac{1}{\nu } t [ l_0/\xi_0]^{1/\nu } h'(t [ l_0/\xi_0]^{1/\nu } ) \right ] \nonumber \\ & = & k_b t l_0^{-3 } \sigma(t [ l_0/\xi_0]^{1/\nu})\end{aligned}\ ] ] where @xmath48 this relation is well known and can be found e.g. in @xcite . we like to emphasis , that it is at the very heart of finite size scaling that the behaviour of the casimir force , which gives the reaction of the film with respect to a change of the thickness and the excess specific heat which gives the reaction of the film with respect to a change of the temperature are given by the same scaling function @xmath2 of the excess free energy . the purpose of the present work is to compute the finite size scaling function @xmath0 by using the energy density of thin films obtained from monte carlo simulations of a lattice model @xcite and by using experimental data for the specific heat of thin films of @xmath10he near the @xmath15-transition @xcite . as a preliminary study , we simulate @xmath13 systems with periodic boundary conditions in all directions . in order to eliminate leading corrections to scaling we have used the improved two component @xmath9 model on the simple cubic lattice . for a precise definition of the model see section [ phi4model ] . using the results obtained for the energy density for a dense grid of temperatures in the neighbourhood of the critical point , we investigate the scaling behaviour and compute the finite size scaling functions @xmath1 and @xmath2 . we find that corrections to scaling are small for the lattices sizes @xmath49 , @xmath50 and @xmath51 that we have simulated . in the case of thin films we analyse data for the energy density that were obtained in @xcite from simulations of the improved two component @xmath9 model on the simple cubic lattice . in @xcite these data where used to compute the specific heat . in order to get a vanishing order parameter as it is observed at the boundaries of @xmath10he films , dirichlet boundary conditions with vanishing field were imposed . in singular quantities these lead to corrections @xmath52 @xcite , which can be expressed by an effective thickness @xmath53 . in @xcite we find @xmath54 for the model that we consider here . note that the boundary conditions also effect the analytic background of the specific heat and the energy density , which also leads to corrections @xmath52 . however it turns out that these corrections are not given by the same @xmath55 as for the singular quantities . taking into account these subtleties we arrive at accurate result for @xmath1 , @xmath2 and @xmath0 . in particular for @xmath0 in the range @xmath56 we find a good match with our previous result @xcite , with previous ones obtained from simulations of the xy model @xcite and experiments on thin films of @xmath10he @xcite ; overall we find a reasonable agreement . ] where we computed the casimir force by taking the derivative of the excess free energy with respect to the thickness @xmath4 of the film . next we compute @xmath0 by using experimental results for the excess specific heat obtained from experiments on thin films of @xmath10he @xcite . even though this is a quite simple exercise , to our knowledge , it has not been done before . for @xmath57 we find a reasonable match with our result @xcite . however in the low temperature phase , for @xmath58 we get results that strongly deviate from @xcite and can be ruled out by plausibility . this corroborates the observation that in the low temperature phase for @xmath58 the excess specific heat does not scale well @xcite . this paper is organized as follows : first we define the model and the observables that we consider . next we discuss the finite size scaling behaviour of the free energy density . in particular , we discuss corrections to scaling caused by dirichlet boundary conditions . in section [ speriodic ] we compute the scaling functions @xmath1 and @xmath2 for @xmath13 systems with periodic boundary conditions . then in section [ sfree ] we compute @xmath1 , @xmath2 and @xmath0 using the data for the energy density of thin films with dirichlet boundary conditions obtained in @xcite . the result for @xmath0 is compared with the one that we @xcite obtained directly from the thermodynamic casimir force . next in section [ sheat ] we compute @xmath0 starting from data for the excess specific heat of films of @xmath10he in the neighbourhood of the @xmath15-transition . finally we summarize and conclude . we study the two component @xmath9 model on the simple cubic lattice . we label the sites of the lattice by @xmath59 . the components of @xmath60 might assume the values @xmath61 . in this work we have performed simulations of lattices with @xmath62 and periodic boundary conditions in all three directions . furthermore we analyse data obtained in @xcite for thin films . in this case lattices of the size @xmath63 and @xmath64 are studied . in 1 and 2-direction periodic boundary conditions and free boundary conditions in 0-direction are employed . this means that the sites with @xmath65 and @xmath66 have only five nearest neighbours . this type of boundary conditions could be interpreted as dirichlet boundary conditions with @xmath67 as value of the field at @xmath68 and @xmath69 . note that viewed this way , the thickness of the film is @xmath70 rather than @xmath4 . this provides a natural explanation of the result @xmath54 obtained in @xcite . the hamiltonian of the two component @xmath9 model , for a vanishing external field , is given by @xmath71\ ] ] where the field variable @xmath72 is a vector with two real components . @xmath73 denotes a pair of nearest neighbour sites on the lattice . the partition function is given by @xmath74 \ , \exp(-{\cal h}).\ ] ] note that following the conventions of our previous work , e.g. @xcite , we have absorbed the inverse temperature @xmath75 into the hamiltonian . in the limit @xmath76 the field variables are fixed to unit length ; i.e. the xy model is recovered . for @xmath77 we get the exactly solvable gaussian model . for @xmath78 the model undergoes a second order phase transition that belongs to the xy universality class . numerically , using monte carlo simulations and high - temperature series expansions , it has been shown that there is a value @xmath79 , where leading corrections to scaling vanish . numerical estimates of @xmath80 given in the literature are @xmath81 @xcite , @xmath82 @xcite and most recently @xmath83 @xcite . the inverse of the critical temperature @xmath84 has been determined accurately for several values of @xmath15 using finite size scaling ( fss ) @xcite . we shall perform our simulations at @xmath85 , since for this value of @xmath15 comprehensive monte carlo studies of the three - dimensional system in the low and the high temperature phase have been performed @xcite . at @xmath85 one gets @xmath86 @xcite . since @xmath85 is not exactly equal to @xmath80 , there are still corrections @xmath87 , although with a small amplitude . in fact , following @xcite , it should be by at least a factor 20 smaller than for the standard xy model . in @xcite we find for @xmath88 by fitting the data for the second moment correlation length in the high temperature phase @xmath89 , where @xmath90 . we shall use this definition of the reduced temperature also in the following discussion of our numerical results ; hence @xmath91 . note that in the high temperature phase there is little difference between @xmath92 and the exponential correlation length @xmath93 which is defined by the asymptotic decay of the two - point correlation function . following @xcite @xmath94 for the thermodynamic limit of the three - dimensional system . hence at the level of precision reached here it does not matter whether @xmath95 or @xmath96 is used in the scaling variable @xmath97^{1/\nu}$ ] . the reduced free energy density is defined as @xmath98 i.e. compared with the free energy density @xmath99 , a factor @xmath100 is skipped . note that in eq . ( [ hamiltonian ] ) @xmath75 does not multiply the second term . therefore , strictly speaking , @xmath75 is not the inverse of @xmath100 . in order to study universal quantities it is not crucial how the transition line in the @xmath75-@xmath15 plane is crossed , as long as this path is not tangent to the transition line . therefore , following computational convenience , we vary @xmath75 at fixed @xmath15 . correspondingly we define the ( internal ) energy density as the derivative of the reduced free energy density with respect to @xmath75 . furthermore , to be consistent with our previous work @xcite , we multiply by @xmath101 : @xmath102 it follows @xmath103 which can be easily determined in monte carlo simulations . from eqs . ( [ fdef1],[edef1 ] ) it follows that the free energy density can be computed as @xmath104 let us briefly discuss the scaling behaviour of the reduced excess free energy per area . since we study an improved model we ignore corrections @xmath105 in the following . we take into account leading corrections due to the boundary conditions by replacing the thickness @xmath4 of the film by @xmath106 at the appropriate places . we split the free energies in singular ( s ) and non - singular ( ns ) parts : @xmath107 where @xmath108\ ] ] is a universal finite size scaling function and @xmath109^{1/\nu}$ ] . following rg theory the non - singular part is not affected by finite size effects . however it is not clear a priori how dirichlet boundary conditions affect the non - singular part of the free energy . therefore we allow for @xmath110 and @xmath111 . taking the derivative with respect to @xmath4 we get the thermodynamic casimir force per area @xcite @xmath112 where @xmath113 note that the boundary terms @xmath114 and @xmath115 do not contribute to the casimir force . first we have studied a finite lattice with @xmath116 with periodic boundary conditions in all directions . this way we avoid corrections caused by the free boundary conditions and possible difficulties related with the kosterlitz - thouless transition of thin films . in our simulations we determine the energy density @xmath117 for the lattice sizes @xmath49 , @xmath50 and @xmath51 for a large number of @xmath75-values in the neighbourhood of the critical temperature . in particular , we have simulated at 159 @xmath75-values in the interval @xmath118 , 205 @xmath75-values in the interval @xmath118 and 161 @xmath75-values in the interval @xmath119 in the case of @xmath49 , @xmath50 and @xmath51 , respectively . for most of these simulations we performed @xmath120 measurements . for each of these measurements we performed one metropolis sweep , two overrelaxation sweeps and a number of single cluster @xcite updates . the number of single cluster updates is chosen such that the number of updates times the average size of a cluster roughly equals @xmath13 . as random number generator we have used the simd - oriented fast mersenne twister algorithm @xcite . in total we have used about 3 weeks of cpu - time on a single core of a quad - core opteron(tm ) 2378 cpu ( 2.4 ghz ) . in order to compute the excess energy density @xmath121 we have used the results for @xmath122 obtained in section 4.1 of @xcite . the reduced excess free energy density behaves as @xmath123 where @xmath124^{1/\nu}$ ] . it follows for the excess energy density @xmath125^{1/\nu } l^{-3 } h'(x ) \;.\ ] ] since we study an improved model and periodic boundary conditions in all directions , we expect that analytic corrections are leading . in figure [ eplot ] we plot @xmath126^{-1/\nu } l^3 [ e(t , l ) - e_{bulk}(t)]$ ] as a function of @xmath127^{-1/\nu}$ ] . the curves for @xmath49 , @xmath50 and @xmath51 fall nicely on top of each other , showing that corrections to scaling are numerically small . the excess energy is positive for all temperatures . at @xmath128 the function assumes a maximum . the value at the maximum is @xmath129 . in the high temperature phase , for increasing @xmath60 the function @xmath1 rapidly approaches zero . in contrast , in the low temperature phase it is only slowly approaching zero with decreasing @xmath60 . this behaviour might be explained by the presence of a goldstone mode in the low temperature phase . next we have computed the excess free energy using eq . ( [ integratef ] ) . to this end we integrated our data for the excess energy by using the trapezoidal rule . we have started the integration at @xmath130 , @xmath131 and @xmath132 for @xmath49 , @xmath50 and @xmath51 , respectively . at these values of @xmath133 the deviation of the excess energy from zero is of similar size as the statistical error . in figure [ fplot ] we plot @xmath134 $ ] as a function of @xmath127^{-1/\nu}$ ] . as one should expect , the curves for @xmath49 , @xmath50 and @xmath51 fall nicely on top of each other . since the excess energy is positive for all values of @xmath60 , the excess free energy is monotonically increasing with increasing @xmath60 . at the critical point the scaling functions assumes the value @xmath135 . in fig . 6 of @xcite results for an isotropic @xmath13 system with periodic boundary conditions for the ising universality class and the limit @xmath136 of @xmath137 symmetric models are given . the ising result is obtained from a perturbative approach in three dimensions fixed , while the @xmath136 result is exact . in the ising case , @xmath2 shows a minimum in the low temperature phase , while for @xmath136 it is monotonically decreasing as the temperature decreases . hence qualitatively the behaviour for @xmath138 , studied here , is the same as for @xmath136 . here we study thin films with free boundary conditions in the short direction . this geometry is relevant for the comparison with experimental results obtained for thin films of @xmath10he . most of the monte carlo data are taken from our previous work @xcite , where we have simulated films of the thicknesses @xmath139 , @xmath50 and @xmath51 . therefore we refrain from giving the details of the simulations and refer the reader to @xcite . analogous to the previous section we compute the scaling function @xmath1 of the excess energy density and @xmath2 of the excess free energy density . using these we obtain the scaling function @xmath140 of the thermodynamic casimir force . in @xcite we have taken great care to get the deviations of the energy density from its effectively two dimensional thermodynamic limit under control . in order to achieve this , quite large ratios @xmath141 are needed in the neighbourhood of the peak of the specific heat . in the case of @xmath139 we have simulated lattices of a size up to @xmath142 , and for @xmath143 up to @xmath144 . for @xmath145 we have skipped the interval @xmath146 , since we could not simulate sufficiently large values of @xmath147 . in figure [ naivplot ] , similar to the previous section , we have plotted @xmath148^{-1/\nu}$ ] versus @xmath149^{1/\nu}$ ] , where now @xmath150\ ] ] is the excess energy per area . in contrast to the previous section we find that there is a huge discrepancy between the three curves . the dominant effect seems that the curves are shifted by a constant with respect to each other . note that replacing @xmath4 by @xmath151 with @xmath54 @xcite does change this situation only little . in section [ theory ] we have argued that the analytic background of the energy density might suffer from a boundary correction that is not given by the effective thickness @xmath151 that describes the leading boundary corrections of singular quantities . below we shall study this question in detail at the critical point of the bulk system , where we have data for thicknesses up to @xmath152 available . in order to get a better understanding of the corrections we have studied in detail the behaviour at the critical point of the three dimensional bulk system . in the context of @xcite we have simulated lattices of the thickness @xmath153 , @xmath154 , @xmath50 , @xmath155 , @xmath51 , @xmath156 and @xmath157 and @xmath158 . in @xcite we have checked that this choice of @xmath159 , @xmath160 is sufficient to approximate well the effectively two dimensional thermodynamic limit of the film at the critical point of the three dimensional system . our results for the energy density are summarized in table [ ene6 ] . ._[ene6 ] energy density of thin films of the thickness @xmath4 at the inverse critical temperature @xmath86 of the three dimensional systems . in all cases _ [ cols=">,<",options="header " , ] in the case of periodic boundary conditions in all directions , the energy density at the critical point behaves as @xmath161 where @xmath45 is the dimension of the system . using lattices of the size @xmath62 with @xmath4 up to @xmath162 we find @xcite @xmath163 in order to take into account corrections due to the free boundary conditions of the thin film we use the ansatz @xmath164 to fit the data given in table [ ene6 ] . as input we have used @xmath165 @xcite , @xmath166 given in eq . ( [ periodicebc ] ) and @xmath54 @xcite . the parameters of the fit are @xmath167 and @xmath168 . fitting all data with @xmath169 we get an acceptable @xmath170d.o.f . we find @xmath171 when fixing @xmath172 and @xmath173 fixing @xmath174 . hence @xmath167 is clearly different from @xmath175 and it shows little dependence on the value taken for @xmath175 . we have checked that the error of @xmath167 due to the uncertainties of @xmath6 and @xmath166 can be ignored . the boundary correction @xmath176 should be an analytic function of the reduced temperature . in a first attempt we shall approximate it by its value at the critical point of the three dimensional system found above . hence in figure [ scaleplot ] we plot @xmath177^{-1/\nu } $ ] where @xmath178 as a function of @xmath109^{1/\nu}$ ] . now we see a quite good matching of the three curves . only for small @xmath60 discrepancies are visible . next we have computed the finite size scaling function @xmath8 of the casimir force following eq . ( [ master ] ) . in the case of @xmath139 and @xmath143 we have used the function @xmath1 as given in figure [ scaleplot ] . in the case of @xmath145 we have taken the missing part in the range @xmath179 from the results for @xmath143 . to this end we have matched the values of the function at @xmath180 and @xmath181 resulting in @xmath182 for @xmath179 . we have computed the function @xmath2 by numerically integrating @xmath1 using the trapezoidal rule . for sufficiently large @xmath60 the casimir force vanishes and therefore @xmath183 . hence for large @xmath60 : @xmath184 in @xcite we found that the thermodynamic casimir force is of similar size or smaller than the numerical errors that we achieve for @xmath185 . we have checked that in this range the scaling function @xmath1 of the excess energy indeed follows eq . ( [ high ] ) . hence we have started the numerical integration in the high temperature phase at @xmath186 with the starting value @xmath187 . in order to check the reliability of our result , we have redone the integration using a set of data points , where we have skipped every second value of @xmath75 . we found an agreement within the statistical errors . our results for @xmath8 are plotted in figure [ casimir1plot ] . in the range @xmath188 the curves obtained from the different values of @xmath4 match quite well . there is also a good match with @xmath8 obtained in @xcite . in particular the value and the position of the minimum of @xmath8 are completely consistent . however for @xmath189 the difference between the curves becomes clearly visible and increases with decreasing @xmath60 . for small @xmath60 , even for @xmath145 there is a huge discrepancy with the result of @xcite . since these discrepancies appear for large values of @xmath190 it is likely that they are mainly caused by analytic corrections . to check this explicitly , we allowed for two different types of corrections : @xmath191^{1/\nu}\ ] ] and for a temperature dependence of the boundary correction of the analytic part of the energy @xmath192 we find that the curves for @xmath1 obtained from @xmath143 and @xmath145 can be nicely matched by adjusting the two parameters @xmath193 and @xmath194 . matching in the interval @xmath195 we find @xmath196 and @xmath197 and for the interval @xmath198 @xmath199 and @xmath200 . using the corresponding results for @xmath1 we have computed the finite size scaling function @xmath0 that is plotted in figure [ casimirmatchplot ] . now we see that the rage of the matching with our previous result @xcite is extended to @xmath201 in the case of the matching range @xmath202 . for still smaller values of @xmath60 discrepancies rapidly increase . likely higher order analytic corrections are the main reason for this behaviour . however also other types of corrections like @xmath203 with @xmath204 @xcite should be taken into account . therefore we abstain from fitting with @xmath205 corrections . in a number of experiments the excess specific heat of thin films of @xmath10he and @xmath37he-@xmath10he mixtures has been measured in the neighbourhood of the @xmath15-transition @xcite . in these works the scaling function @xmath12 which is defined by @xmath206 is extracted from experimental data for the specific heat of the three - dimensional bulk system @xmath207 and of thin films @xmath40 , where @xmath4 is the thickness of the film . since the specific heat is the derivative of the energy density with respect to the temperature , @xmath41 is , up to a constant factor , equal to @xmath11 . to compute this factor let us start from the excess reduced free energy density : @xmath208 where @xmath209 and @xmath5^{1/\nu}$ ] . note that here , as long as the free energy density and @xmath210 are measured in the same units , @xmath2 is uniquely defined ; there is no ambiguous factor . the excess energy density is given by the derivative with respect to @xmath211 . hence @xmath212^{1/\nu } \frac{1}{k_b t_{\lambda } } \beta^{-2 } \simeq - l_0 ^ 3 [ l_0/\xi_0]^{1/\nu } k_b t_{\lambda}\ ] ] where we have approximated @xmath213 in the neighbourhood of the @xmath15-transition . the specific heat as defined in the experiments is given by the derivative of the energy density with respect to the temperature . hence @xmath214^{2/\nu } h''(x ) \;.\ ] ] the results for the specific heat of the experiment are given in @xmath215 . these we convert into @xmath216 to get the same units on both sides of equation ( [ richtig ] ) . note that the thickness of the films in @xcite is quoted in @xmath217 . to this end we need the density @xmath218 @xcite of @xmath10he at the @xmath15-transition , the molar weight @xmath219 of @xmath10he and the boltzmann constant @xmath220 . this amounts to the factor @xmath221 . in @xcite the data are given as a function of the reduced temperature @xmath222 , where @xmath223 . in order to plot them as a function of @xmath5^{1/\nu}$ ] we have used @xmath224 which we @xcite have computed from the amplitude of the bulk specific heat of @xmath10he at vapour pressure @xcite and the universal amplitude ratio @xmath225 @xcite . in figure [ heatplot ] we have plotted @xmath226 k_b^{-1 } l_0 ^ 3 [ l_0/\xi_0]^{-2/\nu}\ ] ] as a function of @xmath5^{1/\nu}$ ] . to this end we have used the data given in @xcite for the thicknesses 483 , 1074 , 2113 , 5039 , 6918 and @xmath227 . note that @xmath228 as can be obtained from the results of section 4.3 of @xcite . for @xmath229 the curves obtained from different thicknesses of the film fall reasonably well on top of each other . it has been noticed @xcite that for @xmath58 , in particular in the neighbourhood of the minimum , the results obtained for different thicknesses differ by quite large factors . in @xcite we have computed the specific heat and the scaling function @xmath12 starting from the data for the energy density discussed above . we find that for @xmath58 our final result is clearly smaller than the experimental ones @xcite . starting from the results for the finite size scaling function @xmath11 obtained from different thicknesses we have computed the scaling function @xmath1 . to this end , we have applied the trapezoidal rule . similar to the previous section , we have started the integration at @xmath186 in the high temperature phase . as starting value we have taken @xmath230 . again we have integrated @xmath1 using the trapezoidal rule to get @xmath2 . here we have taken the same value for @xmath231 as above and @xmath232 . using these results for @xmath1 and @xmath2 we have computed @xmath0 which we have plotted in figure [ experimentalm ] . in order to check the effect of errors due to the finite step - size of the integration , we have repeated the integration , skipping every second value of @xmath60 . in order to check the effect of the singularity of @xmath1 we have fitted the data for the specific heat in the neighbourhood of the transition with the ansatz @xmath233 then we have integrated the ansatz in the neighbourhood of the transition and compared it with the corresponding result from the trapezoidal rule . we find that the numerical results only change little and the conclusions drawn below are not effected . let us now discuss the results that we have obtained : for @xmath234 the curve is monotonically decreasing with decreasing @xmath60 ; in particular no minimum of the function can be observed . for @xmath235 a shallow minimum occurs at @xmath236 ; for @xmath237 the function is decreasing again with decreasing @xmath60 . for @xmath238 we see a clear minimum at @xmath239 ; however the value of the minimum is clearly smaller than the one of @xcite . in the case of @xmath240 and @xmath241 we find a quite good match with our result @xcite down to @xmath242 . for @xmath241 the minimum is located at @xmath243 and the value of the minimum is @xmath244 . for @xmath241 the minimum occurs at @xmath245 and the value of the minimum is @xmath246 . for @xmath240 no data for @xmath247 are available . for @xmath241 the curve is decreasing again for @xmath248 with decreasing @xmath60 . up to here the expectation that with increasing thickness of the film the result converges toward the universal scaling function is fulfilled . however for the largest thicknesses studied , @xmath227 even the worst mismatch is found . the curve is monotonically decreasing with decreasing @xmath60 and even for @xmath249 there is quite large mismatch with our result for @xmath8 @xcite . playing around with the data , one finds that smaller values of @xmath11 for @xmath58 are needed to avoid that @xmath8 is decreasing with decreasing @xmath60 for @xmath58 . one should note that eq . ( [ important ] ) only holds for the scaling limit . therefore the observations made here are not an indication that the experimental data are effected by an error . they can also be explained by corrections to the scaling behaviour . this is at least true for the smaller thicknesses like @xmath234 . the result for @xmath227 is more puzzling . one should note that the analysis presented in this section does not depend on the type of boundary conditions that is realized in the experiment . but we think that the behaviour of @xmath0 for @xmath58 obtained here can not be explained by different boundary conditions from those used in the study of the lattice model @xcite . we have studied the relation of the excess specific heat , the excess energy and the thermodynamic casimir force in thin films in the three - dimensional xy universality class . to this end we have exploited the relation ( [ important ] ) @xmath250 among the finite size scaling functions @xmath0 of the thermodynamic casimir force , @xmath2 of the excess free energy and @xmath1 of the excess energy . we have analysed data obtained for the energy density of the improved two - component @xmath9 model on the simple cubic lattice and experimental results for the specific heat of thin films of @xmath10he near the @xmath15-transition @xcite . as a first exercise we have computed the functions @xmath1 and @xmath2 for a finite lattice of the size @xmath13 with periodic boundary conditions in all directions . this way we avoid potential problems related with the kosterlitz - thouless phase transition of the thin film and corrections caused by dirichlet boundary conditions . indeed we find a good collapse of the data already for rather small lattice sizes @xmath49 , @xmath50 and @xmath51 . next we have repeated the same exercise for films with free boundary conditions using the data for the excess energy density obtained in @xcite . here we find a huge mismatch between the thicknesses @xmath139 , @xmath50 and @xmath51 . replacing @xmath4 by @xmath151 with @xmath54 @xcite does not remove this discrepancy . we argue that the non - singular part of the energy density suffers from boundary corrections that are not described by the same @xmath175 which accounts for the corrections in singular quantities . indeed , the analysis of the energy density of films up to the thickness @xmath152 at the critical temperature of the three dimensional system results in a @xmath167 for the analytic background of the energy density that is clearly different from @xmath175 . taking into account this result we find a reasonable collapse of the functions obtained from @xmath139 , @xmath50 and @xmath51 in a large range of the scaling variable @xmath60 . computing the casimir force from this result for @xmath1 we find a good collapse for @xmath251 . in this range of @xmath60 we also find a good agreement with the result for @xmath8 that we have obtained in @xcite . note that in particular the minimum of the scaling function @xmath8 is within this range . we confirm the position @xmath252 and the value @xmath253 that we have obtained in @xcite . next we took into account analytic corrects . the coefficients of these corrections were computed by matching the results obtained from @xmath143 and @xmath145 . this way we could extend the range of agreement with our previous result @xcite down to @xmath254 . we have presented a viable alternative to compute the scaling function @xmath8 of the casimir force . the nice agreement with our previous result @xcite gives us further confidence in the correctness of the results . finally we have computed the scaling function @xmath8 using experimental results for the excess specific heat @xcite . here we find a reasonable match with the theoretical results in the range @xmath229 . in particular we find evidence that the minimum of the scaling function @xmath8 is located at @xmath255 which is consistent with our prediction @xcite but slightly larger than @xmath256 @xcite and @xmath257 @xcite , where the thermodynamic casimir force has been determined for films of @xmath10he with thicknesses @xmath258 . in the range @xmath58 the curves obtained from different thicknesses show quite different behaviour . furthermore in this range @xmath0 is decreasing with decreasing @xmath60 . this behaviour corresponds to too large values of @xmath11 in the range @xmath58 . we think that understanding this problem requires a detailed knowledge of the experiments and is therefore beyond the scope of the present work . this work was supported by the dfg under the grant no ha 3150/2 - 1 . campostrini m , hasenbusch m , pelissetto a , and vicari e , _ theoretical estimates of the critical exponents of the superfluid transition in @xmath10he by lattice methods _ , 2006 phys . b * 74 * 144506 [ cond - mat/0605083 ] saito m , _ an application of finite field : design and implementation of 128-bit instruction - based fast pseudorandom number generator _ , phd thesis , dept . of math . , graduate school of science , hiroshima university , advisor : m. matsumoto ; the numerical program and a detailed description can be found at http://www.math.sci.hiroshima-u.ac.jp/~m-mat/mt/sfmt/index.html " lipa j a , swanson d r , nissen j a , geng z k , williamson p r , stricker d a , chui t c p , israelsson u e and larson m , _ specific heat of helium confined to a 57-@xmath261 m planar geometry near the lambda point _ , 2000 , phys . lett . * 84 * 4894

we discuss the relation of the specific heat , the energy density and the thermodynamic casimir effect in the case of thin films in the three dimensional xy universality class . the finite size scaling function @xmath0 of the thermodynamic casimir force can be expressed in terms of the scaling functions @xmath1 and @xmath2 of the excess energy density and the excess free energy density . a priori these quantities depend on the reduced temperature @xmath3 and the thickness @xmath4 of the film . however finite size scaling theory predicts that the scaling functions depend only on the combination @xmath5^{1/\nu}$ ] , where @xmath6 is the critical exponent and @xmath7 the amplitude of the correlation length . we exploit this fact to compute @xmath8 from monte carlo data for the excess energy density of the improved two - component @xmath9 model on the simple cubic lattice with free boundary conditions in the short direction . we repeat this exercise using experimental data for the excess specific heat of @xmath10he films . the finite size scaling behaviour of the excess specific heat is governed by @xmath11 , which is proportional to the scaling function @xmath12 discussed in the literature . we compare our results with previous work , where the casimir force has been computed by taking the derivative of the excess free energy with respect to the thickness of the film . as a preparative study we have also computed the scaling functions @xmath1 and @xmath2 for finite @xmath13 systems with periodic boundary conditions in all directions , where @xmath14 is the linear extension of the system . 0.5 cm 0.5 cm * the specific heat , the energy density and the thermodynamic casimir force in the neighbourhood of the @xmath15-transition * martin hasenbusch 0.3 cm _ institut fr physik , humboldt - universitt zu berlin _ _ newtonstr . 15 , 12489 berlin , germany _ _ e mail : [email protected]_ 0.4 cm * keywords : * @xmath15-transition , classical monte carlo simulation , thin films , finite size scaling , thermodynamic casimir effect

the search for extrasolar planets around main sequence stars has already harvested thousands of candidates over the past decade ( see e.g. * ? ? ? * ) . although the number of confirmed exoplanets is now well over a thousand , no exomoons have yet been confirmed , though the smallest exoplanet detected has twice the mass of the moon @xcite . to date , no detections of extrasolar rocky objects with masses an order of magnitude lower than a lunar mass @xmath5 g have been reported . in this letter , we propose a method to detect objects with masses several orders of magnitude smaller than @xmath3 in close orbits around white dwarfs ( wds ) . wds are ideal for two reasons . first , the small size of a wd compared to that of a main sequence star makes the detection of faint companion objects and their properties much easier @xcite . second , evidence for circumstellar rocky material already exists . the presence of metal enrichment on the surface of many wds suggests accretion of rocky debris , since the sedimentation timescale of wds is often orders of magnitude shorter than their age . the inferred metal accretion rates can not be accounted for by the interstellar medium @xcite . furthermore , at least 30 wds show evidence for dusty disks that are within the roche zone of the host star @xcite . the standard model " ( see reviews by @xcite and @xcite ) for these disks is that they originate from a close - in minor planet that was perturbed by an outer planet into the roche zone , where it disintegrated into a disk @xcite . alternatively , a small fraction of the disks might be accounted for by scrambled comets @xcite , though the metal abundance of most of the dusty disk wds favors tidal disruption of rocky objects . consistency with the dusty disk model places lower limits on the population of rocky objects around wds . @xcite developed a theoretical model where an underlying mass distribution of asteroids gradually accrete onto wds , and concluded that to explain photospheric absorption lines , asteroids of masses @xmath6 are required for typical wds . the minimum total mass required to explain the metal pollution alone in one dbz wd is @xmath7 @xcite , roughly the mass of ceres , the largest asteroid in the asteroid belt . with recent work estimating that @xmath8 wds with an age of 20200 myr accrete planetary debris @xcite , a method to directly detect such objects should either lead to many detections or a revision of the current paradigm . this paper is organized as follows . in 2 , we study the properties of objects with masses much less than @xmath3 arounds wds . in 3 , we calculate the minimum size of objects that could be detected with existing telescopes and the future james webb space telescope ( jwst ) . finally , we discuss possible issues as well as the significance of detection ( or upper limits ) in 4 , and summarize our conclusions in 5 . the tidal locking timescale for a gravitationally bound rocky object with @xmath10 at an orbital radii @xmath11 au around a wd is @xmath12 yr @xcite , i.e. , much shorter than the planet formation timescale of @xmath13 10 myr for the objects under consideration . we thus expect that virtually all asteroids should be tidally locked to the host wd . over the age of the solar system , only objects with a mass @xmath14 can retain volatiles at 50 k @xcite . at higher temperatures , a much larger mass is needed , since the timescale for the atmosphere to leak away is dominated by an exponential in inverse temperature , @xmath15 , where @xmath16 is the atomic weight and @xmath17 is the radius of the object which has units of pressure : @xmath18 . with @xmath19 , we obtain a critical mass for sphericity @xmath20 . for a rocky body with a composition similar to that of the moon , @xmath21 kpa and @xmath22 . thus we expect objects with @xmath23 to be spherical . a more detailed analysis @xcite refines @xmath24 by less than an order of magnitude . ] . therefore , the objects under consideration here are unlikely to have atmospheres . in the absence of an atmosphere , the temperature differences between the night and day sides of an object must be of order unity , because thermal conduction is inefficient . the situation is more dramatic than in the case of our moon , because one side of the object will be tidally locked with the wd illumination . the power transmitted by thermal conduction is @xmath25 whereas the power radiated due to blackbody radiation is of order @xmath26 for an object at 600 k with mass @xmath4 and a thermal conductivity @xmath27 , radiation is 7 orders of magnitude more efficient than conduction . another way to decrease the contrast between the day and night side is by tidal heating of the objects . the power @xmath28 generated by the oscillating tidal forces as the object cycles from apoapsis to periapsis is @xcite @xmath29 where @xmath30 is the minor planet s tidal dissipation function @xcite . for @xmath31 and a semi - major axis @xmath32 au , @xmath28 is 15 orders of magnitude less than @xmath33 . the upper bound on the temperature of the object is obtained at the roche radius @xmath34 , giving @xmath35 where @xmath36 , @xmath37 , and @xmath38 are the temperature , radius , and mass of the wd . dust and small rocky objects are expected to disintegrate at the rock sublimation temperature of @xmath39 1500 k @xcite . motivated by dusty debris observations , we target asteroids which do not contain a significant amount of ice , since these will sublimate at even lower temperatures . that magellan fire or jwst miri could detect for both sirius b and g29 - 38 with a @xmath13 hr exposure as a function of the minor planet temperature @xmath40 . on the upper axis , we plot the normalized orbital period @xmath41 of the asteroid . for sirius b , @xmath42 and for g29 - 38 , ground based telescopes like magellan are only sensitive out to a wavelength of @xmath44 , and thus quickly lose sensitivity to cooler asteroids . limited photometric stability of jwst may increase @xmath45 . ] next , we show that it is feasible to observe objects with a mass as low as @xmath4 with jwst . the temperature @xmath40 of a minor planet with exposed area @xmath46 separated by a distance @xmath47 from a wd with temperature @xmath36 is given by balancing the heating from the wd illumination with cooling from blackbody radiation : @xmath48 , where we neglect the albedo @xmath49 for a rocky object . this yields @xmath50 . at the roche zone orbital separation of @xmath51 au , @xmath40 could reach @xmath52 . for a fixed orbital distance , the signal - to - noise ratio @xmath53 depends on the flux of the wd as follows : @xmath54 where @xmath55 is the distance from earth to the wd . the distribution of wd temperatures can be written as @xmath56 . assuming self - similarity in the temperature profile as the wd cools , @xmath57 . for a constant production rate of wds , @xmath58 , this yields @xmath59 , which is in reasonable agreement with observed wd statistics @xcite . since the number of wds out to a distance @xmath55 scales as @xmath60 , this implies that the minimum distance needed to observe a white dwarf with temperature @xmath61 scales like @xmath62 , or @xmath63 therefore it is typically advantageous to target hot wds , even though they are rarer and therefore farther away . note that equation ( [ snt ] ) applies to asteroids whose temperatures are roche limited . in the rayleigh - jeans part of the spectrum , the flux from the asteroid will oscillate with an amplitude @xmath64 , where @xmath65 is the inclination of the orbital plane relative to the line - of - sight . unless @xmath66 , we will see oscillations with a period equal to the orbital period of the object , as the body alternates between exposing its day to night sides . roughly 100 times more wds are feasible for this technique compared to transiting systems . therefore , we can target wds which are closer by a factor @xmath67 than in transit surveys , with a flux higher by @xmath68 , improving the resulting @xmath53 @xcite . the most abundant wds are a few gyr old and have temperatures in the @xmath69 k range . however , emission from asteroids would be difficult to observe from the ground for cold wds , since the sky is bright at ir wavelengths @xmath70 , whereas the peak emission from a rocky asteroid is at @xmath71 . for instruments sensitive to @xmath72 , only a few systems are hot enough and close enough to be observed from the ground . the brightest wd in the sky is sirius b , with v magnitude 8.3 and an effective temperature of @xmath73 k. the estimated age of sirius b , @xmath74 myr @xcite , is sufficient to allow for planetesimal formation , a process which occurred in our solar system on a @xmath13 10 myr time scale ( see e.g. * ? ? ? although sirius b orbits the main sequence star sirius a , they are separated by 3 11.5 arcsec ( 8.2 31.5 au ) , making it possible to resolve sirius b alone . with a @xmath75 duty cycle over @xmath76 hrs ( for a few hours of total integration time ) , it will be possible to detect the variable emission from an object with a mass @xmath77 with the folded - port infrared echellette ( fire ) on one of the twin 6.5 meter magellan telescopes . of these objects requires a photometric stability of @xmath78 ppm on a few hours timescale . rms fluctuations of @xmath79 ppm have already been achieved on one minute timescales ( see figure 6 of @xcite ) ; fluctuations on 2 hour long timescales should be @xmath80 assuming white temporal noise , giving us the desired sensitivity . however , atmospheric effects such as a time - dependent airmass could lead to systematic errors on hour - long time scales , though calibration against nearby stars may alleviate such issues . although we simulate observations with fire , an imager such as the mmt and magellan infrared spectrograph detection ( mmirs ) would achieve virtually identical limits and allow for real - time photometric calibration against other stars . alternatively , the odds of finding an object may be increased around dusty , metal - rich wds . the most well studied dusty wd is g29 - 38 ( see e.g. @xcite ) which has a temperature @xmath81 k and is 14 pc away . g29 - 38 s roche zone coincides with its sublimation zone , implying that periodic flux variations can occur over exceedingly short timescales @xmath82 hr . in figure 1 , we show the minimum mass @xmath45 that a @xmath13 hr observation with fire on one of the magellan telescopes could detect on both sirius b and g29 - 38 . as the temperature of the objects drops , the minimum mass increases dramatically since fire is only sensitive out to a wavelength of @xmath44 . for g29 - 38 , the required photometric stability is only @xmath79 ppm , as the white dwarf is cooler than sirius b and poisson - limited size of the objects is much larger . in calculating detectability thresholds , we assumed that the signal to noise for detecting an asteroid is limited by the poisson noise of photon counts from the host wd : @xmath83 where @xmath84 is the number of photons per unit wavelength per unit emitted surface area received by the telescope from a blackbody at temperature @xmath61 over an integration time @xmath85 . @xmath86 and @xmath87 are the effective surface areas of the planet and white dwarfs , respectively . the normalization of @xmath88 is fixed by the brightness of a wd . requiring @xmath89 , we solve for the minimum size of a asteroid that can be detected as a function of wd parameters . for real observations , the required exposure time may be greater by a factor of order unity depending on the spacing of the observations in the time domain . once evidence is found for a asteroid , follow up observations with a similar integration time but with a cadence optimized for the period of the asteroid would give the best statistics . submillimeter observations with ground - based experiments like the atacama large millimeter / submillimeter array ( alma ) could constrain the population of cooler asteroids , though the signals would be periodic on a longer timescale as the orbital period scales @xmath90 . hotter asteroids could also be targeted , but alma would then probe the rayleigh - jeans part of the spectrum where it is not possible to constrain the temperature of the asteroid and therefore difficult to distinguish it from atmospheric features on the surface of a wd ( see 4 ) . a space - based mid - infrared telescope will dramatically improve our constraints on rocky bodies , allowing detection of asteroids around older , cooler wds . in particular , the mid infrared instrument ( miri ) onboard jwst is sensitive in the 5 28.3 @xmath91 spectral range , which will dramatically decrease the minimum observable mass , especially at lower temperatures where the peak emission is at @xmath92 . in figure 1 , we show the minimum mass that a @xmath13 hr observation with jwst could detect on both sirius b and g29 - 38 . note that the logarithmic slope is much shallower for jwst than magellan . at @xmath93 k , jwst will improve constraints on asteroid masses by three orders of magnitude if a photometric stability of @xmath94 ppm per one - minute exposure is achieved . for cooler , more distant white dwarfs , the required photometric stability is less stringent . a potential source of error in future observations might be distinguishing asteroids from atmospheric features on the wd such as spots , although there is currently no evidence @xcite . one simple way to break the degeneracy is to constrain the effective temperature of the object . in particular , the period of the flux oscillation allows one to predict ( up to a factor of order unity due to uncertainty in albedo ) the temperature of the asteroid . for a fixed integration time , the fractional uncertainty in the temperature of the minor planet is @xmath95 in the limit where the spectroscopic band is much wider than the wavelength spread in the emitted spectrum of the asteroid . unless experimental constraints force us to observe only in the rayleigh - jeans regime of the spectrum , only a factor of order unity more integration time will be necessary to constrain the asteroid s temperature to a value much cooler than any plausible spot on the wd surface . if multiple asteroids are orbiting close in , fourier analysis in the time domain should distinguish the various objects , as asteroids on different orbits will have different periods . note that the discovery of objects at multiple periods would require be increasingly harder to explain with spots , as a naive model would have all spots rotate with the same period . furthermore , detection of multiple blackbody components with effective temperatures given by the inferred orbital radii would be a way to confirm the multiple asteroid interpretation . on the other hand , if there are multiple asteroids with very similar periods but different orbital phases , @xmath53 will be reduced as the differing phases wash out the amplitude of the periodic signal . in the limit of a thin dusty disk , the observable periodicity is lost . consequently , our observational strategy is optimal if asteroids do not spend too long near the roche zone ( such that many asteroids could be in similar orbits ) nor too little time near the roche zone ( such that few asteroids will ever be detected ) . the detection of asteroids around wds would be a powerful probe of planetary physics . a survey of nearby wds by jwst could constrain the distribution function @xmath96 , where @xmath40 and @xmath97 are the temperature and mass of asteroids , respectively . @xmath98 can then be used to infer the statistical evolution of asteroids with time , since @xmath36 is linked to the age of the wd by the wd cooling function . for example , the tidal disruption model implies that @xmath99 must be intimately tied to the abundance of dusty disks @xmath100 . a generic prediction of this model is that any changes in the mass of debris disks should be accompanied by metal accretion onto the surface of the wd and/or a non - trivial @xmath101 dependence . since constraints have already been placed on @xmath102 , observations of @xmath99 could falsify or refine this model . in addition to detecting minor planets , this technique could also harvest exoplanet detections , though we would expect @xmath103 for planets with atmospheres . to distinguish a large planet with a small @xmath104 from a small object with a larger @xmath104 , the blackbody spectrum of the object could be obtained and used to infer the effective surface area . furthermore , the mass of close - in exoplanets could in principle be constrained by observations , as tidal forces should pull the planet into an ellipsoid . as a planet orbits the wd , the effective area of the planet will change over a period @xmath105 , as opposed to flux oscillations with a period @xmath61 from the planet s day / night temperature gradient . + we have shown that the periodic flux variability from the night / day oscillations of close - in asteroids allows existing and future telescopes to detect objects with a small fraction of the moon s mass . a possible strategy would be to use an existing - ground based telescope to search for asteroids around nearby wds with @xmath106 12,000 k. jwst can then be used to follow up ground - based observations in order to improve the associated constraints by a few orders of magnitude on already - observed targets and to sample the asteroid environments around the entire population of local wds . the unique environment of wds thus affords an unprecedented window into the physics of planet formation .

since white dwarfs are small , the contrast between the thermal emission of an orbiting object and a white dwarf is dramatically enhanced compared to a main sequence host . furthermore , rocky objects much smaller than the moon have no atmospheres and are tidally locked to the white dwarf . we show that this leads to temperature contrasts between their day and night side of order unity that should lead to temporal variations in infrared flux over an orbital period of @xmath0 to @xmath1 days . ground based telescopes could detect objects with a mass as small as @xmath2 of the lunar mass @xmath3 around sirius b with a few hours of exposure . the james webb space telescope ( jwst ) may be able to detect objects as small as @xmath4 around most nearby white dwarfs . the tightest constraints will typically be placed on 12,000 k white dwarfs , whose roche zone coincides with the dust sublimation zone . constraining the abundance of minor planets around white dwarfs as a function of their surface temperatures ( and therefore age ) provides a novel probe for the physics of planetary formation

the difficulty in establishing a sensible global solution to the problem of low @xmath0 ( reynolds number ) viscous flow around simple objects , where the flow is uniform at infinity , has fascinated applied mathematicians for just over 150 years . stokes ( 1851 ) established that there was no solution to the two - dimensional , steady , incompressible , equations for asymptotically uniform flow around a cylinder in the biharmonic limit . this situation is now routinely described in modern tutorial discussions . see , for example , chapter 7 of acheson ( 1990 ) for an exercise on stokes paradox and a discussion of the corresponding situation for the sphere . the biharmonic equation can be solved in a neighbourhood of the cylinder ( i.e. a circle ) by any stream function of the form @xmath1 but there is no choice of @xmath2 for which @xmath3 for large @xmath4 . the analysis of this problem has lead to several classic papers ( oseen , 1910 ; lamb , 1911 ) and its understanding through the use of matched asymptotic expansions ( mae ) is one of the triumphs of perturbation theory . the reader is referred to van dyke ( 1964 ) for his classic survey of the work of kaplun ( 1957 ) , proudman and pearson ( 1957 ) and other key references . however , the mae approach , despite its immense power and diversity of expanding applications , does not give a clean resolution of the original difficulty in that such methods rely essentially on computing and then matching solutions to the problem defined on two regions : close to the cylinder and far from the cylinder . the purpose of this paper is to address the problem of finding _ global _ solutions for the low reynolds number limit , i.e. to resolve the original paradox . there are , of course , other approaches to the paradox . recently , villas boas @xcite has considered the problem from a three - dimensional perspective and points out that there is then no paradox . a large class of fluids can be characterized by their density , @xmath5 , a scalar field not presumed to be constant , and their dynamic viscosity @xmath6 . the flow is characterized by a velocity vector field @xmath7 , and an associated scalar pressure field @xmath8 . conservation of mass is expressed by the continuity equation @xmath9 and the conservation of momentum is expressed by the navier - stokes equations acting on _ vectors _ should be understood as the ordinary laplacian acting on cartesian components . ] @xmath10 if the fluid is incompressible in the sense that @xmath5 is a constant in both time and space , we have the condition : @xmath11 to analyze matters further , we introduce the vorticity vector @xmath12 in the following discussion we demand incompressibility but allow for non - zero vorticity . using simple identities from vector calculus the navier - stokes equations may then be recast in the form @xmath13 taking the curl of this , we arrive at the vorticity equation @xmath14 where the kinematic viscosity @xmath15 . since the velocity field is divergence - free , we may introduce a vector potential @xmath16 such that @xmath17 and furthermore we may choose it so that it is divergence free : @xmath18 the vector potential can be reduced to a single function when there is an appropriate symmetry . the resulting object is a stream function . for example , planar 2d flow is obtained by setting ( and note that this automatically satisfies satisfies the divergence condition ) @xmath19 next we note that under the assumption that @xmath20 satisfies @xmath21 @xmath22 and the vorticity equation becomes @xmath23 for problems where it is possible to identify a natural length scale @xmath24 and a natural speed @xmath25 , it is normal practice to perform a non - dimensionalization of the variables and introduce the reynolds number @xmath26 . then the navierstokes equation becomes ( after rescaling the variables suitably to @xmath27 ) : @xmath28 throughout this paper we shall work in units in which the _ radius _ @xmath29 of the cylinder is taken to be unity . in such units the reynolds number is here based on the radius on the diameter . ] and is given by @xmath30 . the two stream functions are related by @xmath31 and furthermore @xmath32 etc . the old historical approach to the limiting case when @xmath33 and the flow is time - independent is to take the view that the non - linearities may then be ignored ( provided the non - linear term is well behaved ) and the time - independent navierstokes equations reduce to @xmath34 which is the biharmonic limit , also known as stokes flow . it is now well known ( see for example , chapter 8 of van dyke ( 1964 ) ) that the neglect of the non - linear terms can lead to inconsistencies , as is evidenced by the lack of any solution for asymptotically uniform two - dimensional flow past a cylinder . here we introduce the scalar vorticity function @xmath35 and write the navierstokes as the pair @xmath36 if we had exponential growth in vorticity , @xmath37 , then the linearized form would be @xmath38 or in terms of a single condition : @xmath39 which is the laplace - helmholtz equation . let us focus temporarily on the linearized time - dependent case based on the laplace - helmholtz equation . we let @xmath40 . it is an elementary exercise ( and we shall give equivalent details in a different context later ) to establish that for all @xmath41 , the laplace - helmholtz equation @xmath42 has a solution of the form @xmath43 \sin ( \theta ) \label{tdsf}\ ] ] that satisfies @xmath44 in other words , no matter how small the value of @xmath6 , the paradox does not exist . to summarize : if we have some time - dependence in the simplified form of exponential growth in vorticity the linearization is precisely the laplace - helmholtz problem rather than the biharmonic problem . addition of this type of extra variability , no matter how small , shows that the paradox evaporates , i.e. the existence of the paradox is unstable . villas boas @xcite , who has considered the problem from a three - dimensional perspective , also demonstrates that the paradox evaporates when extra _ spatial _ variability is incorporated . in subsequent analysis we will use the laplace - helmholtz model as a approximation for the time-_independent _ 2d case , but where the modification is regarded as a simple way as approximating the combined effect of the non - linearities via a linear term . the question is as to whether we can do this in a sensible and self - consistent manner . we shall look at this in a variety of ways , starting with a rather _ ad hoc _ approach . the production of a global solution to the equations requires a slightly unusual approach . we take the time - independent equation @xmath45 as our starting point and note some obvious facts . first , if we have a @xmath46 satisfying laplace s equation , then it satisfies equation eq . ( [ oldfour ] ) identically . second , in the high reynolds number limit , although this is a singular limit , we note that the jacobian term should then vanish identically , so that any solution of @xmath47 satisfies eq . ( [ oldfour ] ) when @xmath48 . this will be true of course in the linear case when @xmath49 and this last equation is of helmholtz type . now let @xmath50 be a ( possibly complex ) number of order 1 and @xmath51 a parameter to be determined . in particular , if we consider solutions , @xmath52 , of a parametrized helmholtz equation in the form @xmath53 then the jacobian will vanish and we note that @xmath54 next suppose that we consider a more general stream function @xmath55 as @xmath56 then a trivial calculation tells us that @xmath57 our approach is therefore to combine solutions for potential flow , @xmath46 , with a solution @xmath52 of the helmholtz equation eq . ( 23 ) that also satisfies the fourth order equation given by eq . ( 24 ) without the identically zero non - linear term : @xmath58 when @xmath33 this pde of course approaches the biharmonic equation provided @xmath59 . then eq . ( [ expandedns ] ) offers the possibility that we can solve a pde of the form : @xmath60 where @xmath61 and @xmath62 can be calculated and analyzed . the introduction of the term of order @xmath63 in the right side of eq . ( [ expandedns ] ) may seem like an artificial device , but given that ( a ) we are seeking solutions of the problem as @xmath33 ; ( b ) we are not modifying higher derivatives in the equations ; ( c ) we have a valid perturbation equation in eq . ( 28 ) , we consider that it is valid to proceed with this approach . the author recognizes that the argument in the previous sub - section is somewhat _ ad hoc _ , and initially gives us no idea how to justify the choice of @xmath50 or @xmath51 , or , as we shall introduce , the relevant composite parameter @xmath64 . however , there is another other route to justifying this approach . let @xmath35 . then the full 2d time - independent incompressible navier stokes equation is precisely @xmath65 we already know from the existence of the stokes paradox that simple linearization by setting @xmath66 is hopeless . we might therefore consider starting from other linearizations , for example , the helmholtz linearization @xmath67 where @xmath64 has to be determined either from experiment or from deeper theoretical considerations . the linear starting point for analysis is then not the biharmonic system but the helmholtz - laplace equation : @xmath68 in principle then we can consider justifying the choice of @xmath64 by matching @xmath69 with @xmath70 as closely as possible under a suitable norm . we will prove in the next section that _ for all @xmath71 the helmholtz - laplace equation admits a solution satisfying both the boundary conditions on the circle and the velocity condition at infinity _ , obtained by considering elementary solutions that are linear combinations of solutions of the laplace equation and solutions of the helmholtz equation . note that this approach and our _ ad hoc _ scheme are linked by the relationship @xmath72 the determination of @xmath50 and @xmath51 are interesting challenges . we shall see in section 8 that there are good theoretical grounds for setting @xmath73 . for now we will leave both parameters general . we now consider the well - trodden route to the analysis of a stream function associated with a flow that is uniform at infinity and satisfies appropriate boundary conditions on @xmath74 . using polar coordinates , we therefore want @xmath75 and @xmath76 we build the solution for @xmath46 and @xmath52 as follows , under the assumption that @xmath46 satisfies the laplace equation and @xmath52 the helmholtz conditions . any sum of the two will satisfy the laplace - helmholtz equation . given that the angular behaviour at infinity is @xmath77 we make the standard assumption and seek solutions @xmath78 the potential flow part ( as usual ) will be taken to be @xmath79 for some constant @xmath80 . our analysis will differ from stokes classic ( stokes , 1851 ) treatment in that @xmath81 does not satisfy the separated biharmonic equation . instead we use the helmholtz equation , which in separated form is just @xmath82 where @xmath83 . the general solution to this ode is given by @xmath84 where @xmath85 are standard `` modified '' bessel functions . it is now quite clear that we can construct @xmath81 so as the preserve the boundary condition at infinity by setting @xmath86 . the @xmath87 function decays exponentially if @xmath71 . we are left with two arbitrary constants that can be determined by satisfying the boundary conditions on the circle given by eq . ( [ circlebc ] ) . this is a matter of elementary algebra using some standard bessel function identities . the final result for the total stream function can be simplified to : @xmath88 \sin ( \theta ) \label{streamfn}\ ] ] we note some interesting facts about this expression . first , it has the _ right _ behaviour as @xmath89 provided @xmath90 , as the bessel function of @xmath4 decays exponentially . second , if we keep @xmath4 fixed and finite and let @xmath91 we obtain @xmath92 and we recover a multiple of stokes ( 1851 ) solution satisfying the boundary conditions on the sphere ( but not at infinity ) , and , furthermore , the multiple is now reminiscent of that arising from the methods of matched asymptotic expansions . the claim therefore is that it is eq . ( [ streamfn ] ) that essentially resolves the paradox , as all the boundary conditions are satisfied , but the low reynolds number limit for finite @xmath4 does not . the third observation is that in the neighbourhood of the surface of the sphere , @xmath93 and that for small @xmath64 this is given by @xmath94 we also note that the limit as @xmath95 is just the potential flow limit : @xmath96 _ mathematica _ code for the evaluation of the stream function and velocity field is given in the appendix . at this stage we have no knowledge about how to fix the parameter @xmath97 , and we now turn our attention to this issue and a more detailed analysis of eq . ( [ expandedns ] ) . we note first that our working assumption of a power dependence of @xmath64 on @xmath0 has not actually been necessary . all we need is that @xmath91 as @xmath33 . we also note that the choice @xmath73 and @xmath98 recovers the result ( van dyke , 1964 ) that the limiting stream function for fixed @xmath4 and small @xmath0 is , from eq . ( [ nearcyl ] ) @xmath99 so that the low @xmath0 limit matches exactly the first term of the mae result . on the diameter we have @xmath100 etc . in the denominator . ] clearly other choices of @xmath50 can be considered , as can other powers of @xmath0 or even a more general @xmath64 still . however , in section 8 we will establish a theoretical basis for estimating these parameters , and argue that @xmath73 . we must now analyze the right side of eq . ( [ expandedns])we call it @xmath101and we do so without any assumption as the the form of @xmath64 . we have , changing variables in the jacobian to polar coordinates : @xmath102\ ] ] recall that we set @xmath103 . having imposed the boundary conditions , we have @xmath104 it follows that @xmath105 where @xmath106 is the reduced jacobian : @xmath107 and @xmath108 denotes the ` helmholtz artifact ' introduced by our approach , and @xmath109 denotes the non - linear jacobian term . some calculation with bessel function identities leads to @xmath110 from which it is manifest that @xmath111 and the non - linear terms vanish on the cylinder boundary irrespective of the choice of @xmath64 . furthermore , the asymptotic behaviour of the bessel functions tells us that provided @xmath71 , @xmath112 as @xmath89 . we deduce that @xmath109 is a bounded function for all @xmath4 . what is its order as a function of @xmath0 ? a straightforward estimate may be given by looking in a neighbourhood of the cylinder . it is straightforward to establish that @xmath113 where the latter expression applies as @xmath114 . we deduce that in the _ immediate _ neighbourhood of the cylinder @xmath115 which is @xmath116 provided only that @xmath114 as @xmath33 . although @xmath106 grows to a maximum as @xmath4 is increased away from unity , before reaching a maximum and then decaying for large @xmath4 , some numerical experiments with _ mathematica _ confirm that the maximum of @xmath117 slowly decreases as @xmath64 decreases to zero . so we can assert that @xmath109 is well behaved and is @xmath116 . when we consider @xmath118 , it is easy to see that @xmath81 has a maximum on @xmath74 , and that @xmath119 as @xmath114 . so this term also behaves . if we desire that the helmholtz artifact tends to zero faster than the non - linear term ( which is desirable for the credibility of our approach ) it is then natural to specify @xmath64 in the form @xmath120 and to demand that @xmath121 . in particular , the choice @xmath73 arranges that @xmath122 and we have established that @xmath123 . it also decays exponentially at infinity . the calculation of the drag can be done by integrating the pointwise force on the cylinder over its surface . the pointwise force has two components . one involves the local rate of strain in the fluid , and the other is the pressure force . the first requires a purely local calculation , but the second requires an integration of the pressure equation from infinity to the cylinder . a question is how this second part can be carried out _ without _ any _ global _ representation of the flow field . for the _ specific _ case of low reynolds number calculations with a certain symmetry there are ways around the problem that we shall exploit presently . to make these matters clear , we shall summarize a first - principles calculation of the force using the results for cylindrical polar coordinates for the rate of strain as given by appendix a of acheson ( 1990 ) . the fluid velocity is given ( in our units ) by @xmath124 on the cylinder the rate - of - strain tensor has components @xmath125 , and @xmath126 the stress tensor @xmath127 , so the force on the cylinder boundary is @xmath128 we project this into the @xmath129-direction and integrate over the circle to get the following formula for the drag @xmath130 ( force per unit length on the cylinder ) : @xmath131d\theta\ ] ] for low @xmath0 we estimate the pressure using the momentum equation in the form @xmath132 and a short calculation gives @xmath133 we need to integrate this in from infinity to a general point on the circle @xmath74 . in general this is awkward without a global form of @xmath55 . but if @xmath55 has the form @xmath134 then we have @xmath135\ \ { \rm where}\ \ { \cal l } [ g(r ) ] \equiv \frac{1}{r}\frac{\partial\ } { \partial r}\left(r\frac{\partial g}{\partial r } \right)- \frac{g}{r^2}\ ] ] and the pressure equation becomes : @xmath136 + \ethvec \sin\theta \frac{\partial\ } { \partial r}{\cal l}[g(r)]\right\}\ ] ] we do the integration from infinity along @xmath137 and then work around the circle @xmath74 . carrying this out we obtain @xmath138|_{r=1}\ ] ] we now evaluate the total integral for the force per unit length to obtain @xmath139-\frac{\partial^2 \!g(r)}{\partial r^2 } \right]\ ] ] the drag coefficient for such a 2d problem is defined following tritton ( 1959 ) as @xmath140 and is now easily seen to be given by @xmath141- \frac{\partial^2\!g(r)}{\partial r^2 } \right|\ ] ] evaluated on @xmath74 . given the vanishing of @xmath142 and @xmath143 this simplifies further to @xmath144 in the case given by the mae approach the stream function is given in the neighbourhood of the cylinder by eq . ( 1 ) and then @xmath145 and @xmath146 are given by @xmath147 in our new model @xmath145 as is given by the radial part of eq . ( [ streamfn ] ) , and this time , using some bessel identities , we obtain @xmath148 we recall now the original assumption that @xmath83 . if we fix @xmath149 but allow @xmath50 to vary the drag coefficient of our model is then @xmath150 we can plot the drag coefficient from our model , assuming that @xmath149 , with various choices for @xmath50 , and compare the results with those of tritton ( 1959 ) , which used an @xmath0 based on the diameter . tritton s data has been converted by halving _ his _ reynolds number . also , in making a comparison with fig . 8.5 of van dyke ( 1964 ) , it appears that van dyke has plotted @xmath151 , as otherwise it is not possible to reconcile that plot with tritton s data . in figure 1 we show the data for the first three fibres used by tritton ( those that consider the lowest @xmath0 ) and * our model with @xmath50 picked to match the mae result , @xmath152 ; * our model with a least squares best fit @xmath153 ; * the one term mae result based on @xmath154 ; * the two term mae result based on @xmath155 . the choice of constants in the mae approach is itself somewhat arbitrary . that made by kaplun ( 1957 ) , giving eq . ( [ nearcyl ] ) , is little more than convenience . these results give support to the new model and the agreement with experiment we get by taking @xmath50 about one half that implied by the mae approach is rather tantalizing . another use for the type of low reynolds numbers solution developed here is to provide a basis for _ iteration _ of the solution in powers of @xmath0 . the presence of the paradox obstructing a global representation in the standard approach makes this impossible . furthermore , we know from a corresponding analysis of the 3d spherical problem that even when the base solution makes sense , even the first iteration may fail : the _ stokes - whitehead _ paradox emerges . so it is of considerable interest to investigate the iteration of the system . in doing so we will keep @xmath64 general initially . we summarize the solution developed thus far . with @xmath55 as given we have , combining equations ( [ expandedns],45 ) and our definitions , @xmath156 where @xmath157 and @xmath158 where these equations also serve to define the radial terms @xmath159 . from now on we will work in terms of these radial functions , so it is useful just to write down their explicit forms : @xmath160 @xmath161 we now attempt to refine the solution by writing the total solution as @xmath162 ideally we would like to arrange that @xmath163 the right side of this is given by the expansion @xmath164 and we write the remaining parts of the navier - stokes equation as @xmath165 we do not know how to solve this full non - linear system , and propose instead to treat a linearized form . the question now is whether to work with the raw system @xmath166 or to propose a helmholtz - damped variation along the lines of our original approach . we also need to understand which non - linear term to treat first . _ @xmath64 is of the form @xmath167 , for example , we know that @xmath109 is the lowest order in @xmath0 : @xmath168 , so we will give priority to killing this lowest order correction . to develop the solution we need to identify the greens s function for the biharmonic operator . in fact we do not need the full form as the angular structure of the right side is simple - we need only look for a pair of appropriate radial green s functions satisfying appropriate boundary conditions on the cylinder and at infinity . to this end we write @xmath169 and require that @xmath170 and @xmath171 we also require that @xmath172 vanishes as @xmath89 and both @xmath172 and its first derivative vanish on @xmath74 . to proceed further we recall the elementary form of the laplacian in 2d cylindrical polar coordinates : @xmath173 so if @xmath62 is of the form @xmath174 then @xmath175 define the @xmath61the radial modal laplacian operator as @xmath176 we now have a pair of odes in the form @xmath177 @xmath178 in a simpler problem we would now construct a pair of green s functions @xmath179 , @xmath180 , with the properties that @xmath181 subject to , _ if it is possible _ , @xmath182 @xmath183 @xmath184 however , there are significant complications . the helmholtz term we have thought of as an modelling artifact has first to be understood as an optimal linear approximation to the full non - linear system . the first steps on this are considered in section 8 . we will eventually have to manage the fact that no @xmath185 function exists satisfying these boundary conditions . however , it is now consistent to proceed to construct @xmath186 , as this ( a ) exists , and ( b ) is needed for the management of the full non - linear system and its solution . we will now focus on the lowest order correction arising from the inertia term , for which @xmath186 is the required radial green s function . the building blocks for this are two different solutions of the ode @xmath187 the solution to the homogeneous problem is of the form @xmath188 to try to build a green s function , we would write down a pair of solutions : @xmath189 the condition at infinity requires that we set @xmath190 , so that @xmath191 the boundary conditions on the cylinder require that @xmath192 so the inner function is given by @xmath193 we now have four unknowns that must be obtained by imposing junction conditions to obtain the delta - function . the junction conditions are : @xmath194 these equations were solved and simplified using _ mathematica _ v7 and the solutions are @xmath195 finally the two parts of the green s function are ( introducing the @xmath129-dependence explicitly ) @xmath196 the solution for @xmath197 is then given formally by @xmath198 this stream function is @xmath199 as @xmath89 so we claim that a secondary paradox of stokes - whitehead type has been avoided . to proceed further we need to better understand the full non - linear system . the task now at hand is to develop a proper theoretical basis for the estimation of the parameter @xmath64 . to this end we must write down the full equations . converting to polar coordinates , the navier stokes equations under consideration are @xmath200 let us assume that @xmath0 is small enough that the flow remains symmetric about the horizontal axis . we may then write a fourier decomposition @xmath201 then with @xmath202 as above , @xmath203 @xmath204 @xmath205\\ \\ & = \frac{r_e}{2 r } \sum_{l=1}^{\infty } \sum_{n=1}^{\infty } \biggl[n[\sin((l - n)\theta)+\sin((l+n)\theta)]\frac{\partial \phi_l}{\partial r}{\cal l}_n \phi_n \\ & \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ - l [ \sin((l+n)\theta)-\sin((l - n)\theta)]\phi_l \frac{\partial \ } { \partial r}{\cal l}_n \phi_n \biggr ] \end{split}\ ] ] doing the fourier analysis gives us @xmath206 \end{split}\ ] ] thus far no approximations have been made , other than to assume the flow remains symmetric about the @xmath129-axis . now we approximate the model by considering only the contributions of the terms with @xmath207 . we then have the simpler , but coupled and non - linear system @xmath208 @xmath209 with our simple form of @xmath210 that is a solution of the laplace - helmholtz condition @xmath211 we have already determined the solution of the eq . ( [ phitwonl]) @xmath212 is just @xmath197 as already given above . in order to generate maximal self - consistency of the linearization we must choose @xmath64 in such a way as to minimize the mismatch between the right sides of eq . ( [ phionenl ] ) and eq . ( [ philin ] ) . ] let us look at the overall scaling behaviour in terms of the reynolds number . we let @xmath213 , so that @xmath214 the required matching is then of the form @xmath215 which strongly supports the scaling behaviour @xmath216 as @xmath33 . that is , we have a theoretical basis for setting @xmath73 . a determination of @xmath50 requires the introduction of a suitable rigorous criteria for minimizing the mismatch between the linear and non - linear forms , and this is under investigation . note that the _ full _ equations , but limited to the first two angular modes , can certainly be written _ without further approximation _ as @xmath217 for some unknown function @xmath218 , and our method can now be properly understood as that of working with some `` average '' value of @xmath218 , and noting that the resulting solution is free of a paradox . note also that writing the right side of this linearized system as a multiple of @xmath219 is not as arbitrary as it might seem , for we know that the right side of the full non - linear system vanishes identically when @xmath220 . that is , this laplace - helmholtz model has a proper theoretical justification , rather than merely being the basis of a convenient interpolation between the boundary conditions on the cylinder and at infinity . we could also consider generalizations where an improved _ ansatz _ for the form of @xmath218 is employed . the optimal average form and improved functional choices are under investigation . only once this has been done would it make sense to consider further iteration . some initial considerations suggest that as a function @xmath221 and that @xmath218 might be asymptotic to a constant independent of @xmath0 as @xmath89 , but further analysis is needed . we have constructed a _ global _ stream function satisfying the two - dimensional viscous incompressible steady equations in the limit @xmath33 . the stream function satisfies the correct boundary conditions on a cylinder and infinity . the `` perturbation '' to the linearized equations introduced to accomplish this is of higher order in @xmath0 than the inertia terms , and this modification is now properly understood as a linear representation of the full non - linear theory . the results suggest improved agreement with experimental data over those obtained by the mae approach . further work is needed on this approach , in particular on comparisons with newer data sets and determination of the remaining free parameter @xmath50 from theoretical considerations . although the approach of this paper initially started with rather _ ad hoc _ considerations , the method developed here is founded on a deeper consideration based on the approximation of the non - linear equations with an optimal linear approximation based on equations of laplace - helmholtz type , rather than the biharmonic equation . the need to properly treat the non - linearities has also been illustrated by the observation that the presence of a paradox is unstable with respect to small changes in the system . the viscous and `` paradox - free '' stream function is given in non - iterated form by @xmath222 \sin ( \theta ) \ ] ] and the parameter @xmath223 . the drag coefficient with tritton s conventions is : @xmath148 the choice of @xmath50 from theoretical considerations has yet to be determined , but experimental drag data suggests @xmath153 . the stream function satisfies the correct boundary conditions on the cylinder and at infinity for all @xmath71 , and also reduces to that for potential flow as @xmath95 . _ mathematica _ code for the stream function and velocity field are given in the appendix . i am grateful to bin zhou for his comments on the earlier ( 2006 ) version of this paper . 10 acheson , d.j . 1990 . _ elementary fluid dynamics_. clarendon press . kaplun , s. 1957 low reynolds number flow past a circular cylinder . _ j. math . mech . _ * 6 * , 595 - 603 . lamb , h , 1911 on the uniform motion of a sphere through a viscous fluid . _ philos . mag . _ [ 6 ] , * 21 * , 112 - 121 . milne - thomson , l.m . 1996 _ theoretical hydrodynamics _ fifth edition , dover publications . ockendon , j.r . , howison , s.d . , lacey , a. & movchan , a. 2003 _ applied partial differential equations _ rev . ed . oxford university press . oseen , c.w . 1910 , _ ark . . fys _ * 6 * , no . 29 . proudman , i. , and pearson , j.r.a . expansions at small reynolds numbers for the flow past a sphere and a circular cylinder . _ j. fluid mech . _ * 2 * , 237 - 262 . stokes , g.g . 1851 on the effect of the internal friction of fluids on the motion of pendulums _ trans . soc . _ * 9 * , part ii , 8 - 106 . tritton , d.j . ( 1959 ) experiments on the flow past a circular cylinder at low reynolds number . _ j. fluid mech . _ * 6 * , 547 - 567 . van dyke , m. 1964 _ perturbation methods in fluid mechanics _ , academic press . villas boas , m.b . , 2007 . stokes paradox ; a new approach ( in portuguese ) presentation at 8th congreso iberoamericano de ingenieria mecnica . the following code may be useful . first the stream function : .... \[psi][r _ , \[theta ] _ , \[epsilon ] _ ] : = sin[\[theta ] ] ( r - ( 1 + 2 besselk[1 , \[epsilon]]/\[epsilon]/besselk[0 , \[epsilon]])/ r + 2 besselk[1 , r \[epsilon]]/\[epsilon]/besselk[0 , \[epsilon ] ] ) .... .... the streamlines are then easily visualized : contourplot[\[psi][sqrt[x^2 + y^2 ] , arctan[x , y ] , 0.01 ] , { x , -10 , 10 } , { y , -5 , 5 } , aspectratio - > 1/2 , contours - > 50 , regionfunction - > function[{x , y } , x^2 + y^2 > = 1 ] , epilog - > circle[{0 , 0 } , 1 ] ] .... .... cartesianvelocity[x _ , y _ , u _ , \[epsilon ] _ ] : = module[{r = sqrt[x^2 + y^2 ] , \[theta ] = arctan[x , y ] , ur , u\[theta ] } , ur = -((u*((-1 + r^2)*\[epsilon]*besselk[0 , \[epsilon ] ] - 2*besselk[1 , \[epsilon ] ] + 2*r*besselk[1 , r*\[epsilon ] ] ) * cos[\[theta]])/(r^2*\[epsilon]*besselk[0 , \[epsilon ] ] ) ) ; u\[theta ] = u*(1 + r^(-2 ) - ( besselk[0 , r*\[epsilon ] ] - ( 2 * besselk[1 , \[epsilon]])/(r^2*\[epsilon ] ) + besselk[2 , r*\[epsilon]])/besselk[0 , \[epsilon ] ] ) * sin[\[theta ] ] ; ur*{cos[\[theta ] ] , sin[\[theta ] ] } + u\[theta]*{-sin[\[theta ] ] , cos[\[theta ] ] } ] .... .... vectorplot [ if[x^2 + y^2 > = 1 , cartesianvelocity[x , y , 1 , 0.1 ] , { 0 , 0 } ] , { x , -10 , 10 } , { y , -5 , 5 } , regionfunction - > function[{x , y } , x^2 + y^2 > = 1 ] , aspectratio - > 1/2 , vectorscale - > small , epilog - > circle[{0 , 0 } , 1 ] ] ....

this paper proposes a solution to stokes paradox for asymptotically uniform viscous flow around a cylinder . the existence of a _ global _ stream function satisfying a perturbative form of the two - dimensional navierstokes equations for low reynolds number is established . this stream function satisfies the appropriate boundary conditions on both the cylinder and at infinity , but nevertheless agrees with stokes original results at finite radius as the reynolds number tends to zero . the navierstokes equations are satisfied to a power - log power of the reynolds number . the drag on the cylinder is calculated from first principles and the free parameter of the approach can be chosen to give good agreement with data on drag . in this revised working paper we put our approach on a firmer mathematical basis using the helmholtz - laplace equation as a linear approximation to the navierstokes system . in so doing we demonstrate the instability of the original paradox . we also demonstrate the absence of a paradox of stokes - whitehead class , and give further theoretical constraints on the free parameters of the model . key words : stokes paradox , fluid dynamics , stokes flow , stream function , biharmonic equation , helmholtz equation , low reynolds number

hydrodynamic simulations of accretion disks in active galaxies ubiquitously predict the presence of mass - outflows originating from the accretion flow . strong evidence for the existence of these outflows comes from the blue - shifted , highly ionized , absorption lines detected in the optical / ultra - violet ( uv)/x - ray spectra of broad absorption line quasi - stellar objects ( bal qsos - _ e.g. _ weymann 1991 , turnshek 1996 ) , however other properties of these outflows , particularly those that we would expect to see imprinted elsewhere on the observed sed , are conspicuous by their absence . few clear signatures of accretion disk winds have been detected , for example , in the observed agn x - ray spectrum ( notable exceptions are the few instances of high ionization iron absorption lines observed in nearby quasars - see section [ 6 - 1 ] ) . the large densities present in the current numerical simulations of line - driven accretion disk outflows predict that these outflows should have a very strong impact on the observed x - ray spectrum of many agn , particularly given the large covering fraction of the high density regions of the flow ( proga , stone & kallman 2000 , proga & kallman 2004 , proga 2005 - see also section [ 4 ] ) . while an origin in an accretion disk wind has often been implicated for the x - ray warm absorber present in many type 1 agn ( _ e.g. _ pg 0844 + 349 & pg1211 + 143 - blustin 2005 , ngc 3783 - kaspi 2002 ) , it is far from clear that this is the correct interpretation for the origin of these features ( see section [ 6 - 4 ] ) . furthermore , the unmodified signatures of compton reflection from the surface of the accretion disk , detected in the x - ray spectrum of many agn , argues against the presence of a considerable column of material above the innermost regions of the accretion disk . the recent development of the numerical code , which simulates uv / x - ray spectra of agn observed through ionized outflowing material , has allowed us to predict and investigate the detailed characteristic spectral features imprinted on the uv / x - ray spectra by an accretion disk wind . this code was originally developed to test the idea of gierliski & done ( 2004 ) that the soft x - ray excess seen in many high accretion rate agn may be due to absorption in an outflow from the disk . the detailed simulations presented in schurch & done ( 2007 - hereafter sd07 ) demonstrated , however , that simple models of radiatively driven accretion disk winds can not reproduce the observed smoothness of the soft excess from a line driven disk wind . such winds instead imprint complex ( sharp ) atomic features on both the transmitted and emitted spectra , the detailed form of which are sensitive to the assumed distribution of gas and its velocity . here we explore the x - ray spectral features produced by a more realistic , self - consistent , line driven accretion disk wind by replacing the assumed gas parameters input to with those derived from hydrodynamic simulations of accretion disk winds . the resulting spectra represent the most realistic picture to date of the direct agn x - ray continuum seen through an accelerating accretion disk wind . in section [ 2 ] we give a brief summary of the key details of the numerical hydrodynamic simulations used in this work . in section [ 3 ] we discuss the modifications made to the original code to allow it to interface with the hydrodynamic simulations . in section [ 4 ] we simulate the uv / x - ray spectra from nine separate lines - of - sight through the hydrodynamic accretion disk wind simulation presented in proga & kallman ( 2004 - hereafter pk04 ) , covering all the main regions of the outflow , and in section [ 5 ] we examine the spectral variability that results from the temporal variability of the simulated hydrodynamic outflow . finally , in section [ 6 ] we discuss the characteristics of the model spectra and make some qualitative comparisons with features observed in current x - ray data . we use the state - of - the - art hydrodynamic accretion disk wind simulations presented in proga & kallman ( 2004 ) to provide a set of self - consistent gas properties for the simulations . the pk04 simulation assumes a 10@xmath0 , black hole accreting at 1.8yr@xmath4 which gives the system a dimensionless accretion rate parameter of 0.5 . this accretes through optically thick material , forming the standard disk spectrum . it is this disk radiation , with temperature varying with radius , which drives the wind . the simulation also includes comptonized continuum radiation from a central source . the continuum source has a luminosity equal to that of the disc , with an effective spectral index of @xmath33 ( so that 90% of the luminosity is emitted in the uv - band and 10% of is emitted as x - rays ) . this component is not included in the line driving calculation for the wind , since the continuum photons typically have higher energies than can be absorbed by the major uv line transitions that dominate the cross - section . it does however have a major effect on the derived photoionization parameter ( see section [ 3 ] ) . the properties of the outflow are calculated over @xmath330 - 1500 r@xmath5 making the wind very geometrically thick . the resulting outflow shows three distinct regions ; a hot , low - density , flow ( @xmath6 ) , a dense , warm , fast equatorial flow ( @xmath7 ) and a thin transition region in - between ( @xmath8 ) . data from the pk04 simulation is recorded in a total of 176 late - time snapshots ( _ i.e. _ snapshots taken once the simulation no longer has a strong dependence on the initial conditions ) . the first 116 snapshots are taken every 10@xmath9s , while the remaining 60 snapshots are taken every 5@xmath1010@xmath9s . the quantities recorded in each snapshot are the gas density , @xmath11 ( g @xmath12 ) , the velocity in three dimensions , v@xmath13 , v@xmath14 & v@xmath15 ( km s@xmath4 ) , and the internal energy , @xmath16 ( ergs ) , for every cell in the simulation . for each cell , @xmath17 , along a radial l.o.s to the central object , we convert these quantities into the input parameters required by , namely , the number density , @xmath18 ( @xmath12 ) , the column density , n@xmath19 ( @xmath20 ) , & the gas temperature , t@xmath21 ( k - see proga , stone & kallman ( 2000 ) , equation 23 ) . is a numerical code that links a series of radiative transfer calculations , incorporating the effects of a global velocity field ( including special relativistic effects ) , to produce model uv / x - ray spectra of agn observed through outflowing material . a detailed discussion of the code is presented in sd07 & sd08 . here we present only a brief discussion of the important characteristics of along with the changes incorporated into the code to allow it to interface with the detailed gas properties from the accretion disk simulations . in a similar fashion to the simulations presented in sd07 , the spectra presented here are calculated using a point - source power - law continuum ( with a spectral index of @xmath22 = 2.4 ) as the initial ionizing spectrum from the central object and assume solar abundances for the wind material . selecting a radial l.o.s through the hydrodynamic simulation data then provides the gas number density ( n@xmath21 ) , temperature ( t@xmath21 ) , column density ( n@xmath19 ) , radius ( r@xmath21 ) and radial velocity ( v@xmath23 ) for each cell , @xmath17 , along the chosen l.o.s . then steps along the l.o.s , using the gas properties from each cell in turn as inputs for each radiative transfer calculation . the ionization state of the material is not given explicitly by the simulation data , however the radial optical depth to every cell in the simulation is known , allowing the ionization state of the material to be computed ( see pk04 ) . for the models presented here , however , we replace the ionization state calculated based on this simple approach with a more rigorous self - consistent treatment which takes energy - dependent absorption and compton scattering in previous cells along the l.o.s into account . the ionization parameter of the gas at each step along a l.o.s is given by @xmath24=l@xmath25/n@xmath21r@xmath26 , where l@xmath25 is the luminosity output from the ( i-1)@xmath27 cell including the effects of compton scattering . we note that this is a substantially different treatment of the ionization parameter than that used in previous calculations . explicitly , in previous calculations , propagates the same spectrum that propagates internally ( l@xmath28 , see sd07 , appendix a1 ) , which does not incorporate the effects of compton scattering ( either in or in ) . uses l@xmath29 to ( amongst other things ) calculate the ionization state for cell @xmath17 . for compton thin columns , ignoring the effects of compton scattering introduces very little error in the calculated ionization parameter , but for compton thick columns , ignoring the effects of compton scattering will result in a considerable overestimation of the true ionization state of the material . the spectral models presented in sd07 & sd08 are all compton thin , however many of the l.o.s through the pk04 simulation have very compton thick total column densities , prompting us to relax our requirement for consistency with and incorporate the effect of compton scattering on l@xmath28 for the calculations presented here . we use the same treatment of compton scattering for l@xmath28 as that used in the calculation of the transmitted x - ray spectrum within ( see sd07 , section 2 ) . we stress that this treatment of the ionization parameter accounts for the energy - dependent absorption ( including the impact of line broadening ) in all the previous cells along the l.o.s , and is , thus , subtly different from the ionization parameter prescription used in pk04 . while the models presented in sd07 & sd08 include both the l.o.s transmitted spectrum and the global emission spectrum , here we present only l.o.s transmitted spectra . currently , calculates global emission spectra assuming that , at each step , the emission originates from a spherically symmetric shell of homogeneous gas with a single , uniform radial outflow velocity and no rotation in either the @xmath30 or @xmath31 directions . clearly this assumption is incompatible with the hydrodynamic wind simulations , which are highly inhomogeneous in density , column and ionization and have strong rotational velocity components . a detailed treatment of the emission spectrum would be extremely computationally intensive and is beyond the scope of this work . a full treatment of the emission spectrum would require a fully relativistic hydrodynamic simulation that also incorporates radiative transfer ; simulations that are beyond our current computational capability ( although considerable development has been made in the field - see , for example , takahashi 2007 ) . the spectra presented here also assume a point - like continuum source , and choosing a purely radial l.o.s to the central source ( _ i.e. _ the simulation origin ) , this allows us to neglect the non - radial outflow velocity component ( _ i.e. _ v@xmath14 & v@xmath15 ) when calculating the transmitted spectrum . however we note that this need not necessarily be the case in agn . for an extended continuum source , the non - radial velocity components will add an additional component of broadening to the line profile . the non - radial velocities may , in fact , dominate the absorption profile if the continuum source is very extended , or in regions of the flow where the radial velocity is small , however we do not expect this to have dramatic consequences for the x - ray spectra presented here for several reasons . firstly , the outflow presented in pk04 is very geometrically thick , suggesting that even if the x - ray continuum source has a considerable angular extent with respect to the innermost regions of the flow , it is unlikely to appear extended from the further reaches of the outflow . secondly , the average non - radial outflow velocities ( v@xmath15 & v@xmath14 ) are of the same order as the average radial velocity , suggesting that , in the case of a marginally extended source , the additional broadening is unlikely to dominate the broadening due to the radial velocity . initially we choose a single late - time snapshot from the pk04 hydrodynamic simulation and use this to investigate the form of the transmitted uv / x - ray spectrum as a function of l.o.s though the simulated wind . we simulate the spectra from nine separate lines - of - sight ; three through the hot , low - density , flow ( @xmath30 = 30@xmath32 , 50@xmath32 & 57@xmath32 ) , two through the dense , warm , fast equatorial flow ( @xmath30 = 67@xmath32 & 75@xmath32 ) and four through the transition region ( @xmath30 = 61@xmath32 , 62@xmath32 , 64@xmath32 & 65@xmath32 ) . table [ columns ] gives the column densities for each l.o.s , while figure [ xscortlos ] shows the calculated spectra from five of lines - of - sight ( for clarity ) , chosen to highlight the dramatic effect that the l.o.s inclination angle has on the observed x - ray spectrum . spectra seen through the hot , low - density , polar flow ( @xmath6 ) remain largely featureless , even though the column can be as large as 1.5@xmath1010@xmath33@xmath20 . this is due primarily to low physical densities in the region ( n@xmath3410@xmath0 @xmath12 ) , which result in an extremely high ionization state throughout the material . while these spectra are not dominated by atomic features , the uv / x - ray photons remain subject to compton scattering which , as highlighted by the 57@xmath32 spectra in figure [ xscortlos ] ( green ) , is important for column densities of @xmath3510@xmath33 @xmath20 . interestingly , the outflow velocities in this region do not exceed @xmath30.1c and are more typically @xmath30.01c . lines - of - sight through the dense , warm , fast equatorial flow ( @xmath7 ) have column densities of @xmath24@xmath1010@xmath36 @xmath20 , which result in almost complete attenuation of the spectrum by compton scattering . densities in this region can be as large as @xmath35@xmath1010@xmath37 @xmath12 , temperatures can fall as low as @xmath310@xmath38 k , however the outflow velocities in this region ( as for the other regions ) do not exceed @xmath30.1c and are more typically @xmath30.01c . the depletion of the spectrum by compton scattering ensures that all but the innermost edge of the l.o.s wind material is essentially neutral ( log@xmath39(@xmath24)@xmath3 - 3 ) , resulting in considerable absorption attenuating the spectrum in addition to the impact of compton scattering itself . the innermost material does , however , effectively shield the outer sections of the wind from direct illumination by the corona source , as required in bal qsos ( murray 1995 ) . for these equatorial l.o.s , the scattered+emitted flux is expected to dominate the uv /x - ray spectrum . sim ( 2008 ) present a detailed discussion of the properties of the scattered+emitted spectrum from a variety of idealized wind models and confirm that high inclination l.o.s through accretion disk winds are expected to be dominated by the scattered+emitted flux ( see figure 3 therein ) . interestingly , they also find that the scattered+emitted spectrum can contain significant , albeit weaker and broader , absorption features , particularly above @xmath36 kev ( see section 6.2 ) . if outflows similar to that simulated in pk04 are common in agn , the large solid angle of the wind ( as seen from the x - ray source ) would imply that @xmath325% of agn should be compton thick , even without the presence of the putative molecular torus commonly proposed by agn unification schemes . the presence of such a wind does not , of course , rule out the presence of the obscuring torus , which , importantly , includes dust and can thus explains the strong ir emission of many compton thick agn . in contrast , lines - of - sight through the transition region of the simulated outflow ( @xmath8 ) result in spectra which have considerable absorption features from ionized species imprinted across the 0.3 - 13 kev range . the physical properties of the material in this region are broadly similar to those found in the hot , low - density , polar flow ( t@xmath4010@xmath41 k , [email protected] ) except that the l.o.s gas density distribution has both a higher maximum ( n@xmath4310@xmath44 @xmath12 ) and , crucially , significantly more high - density cells , than l.o.s through the polar regions . the significantly greater amount of material along the l.o.s ( column densities of [email protected]@xmath1010@xmath46@xmath20 ) means that the majority of this region has an intermediate ionization parameter ( log@xmath39(@xmath24)@xmath33.5 - 4.5 ) and imprints a wide range of atomic features on the spectrum . while the spectra are compton thick and are strongly affected by compton scattering ( factor @xmath310 ) , the scattering is not so dominant that the ionization state of the material is strongly suppressed , or that the scattered+emission spectrum necessarily overwhelms the l.o.s transmitted spectrum ( particularly given the effect of the complex inhomogeneous nature of the flow in the scattered+emission spectrum ) . as well as a strong dependence on inclination , the physical properties of the simulated wind are also strongly time - dependent and this , in turn , leads to considerable variability in the observed uv / x - ray spectrum . figure [ nhvarplot ] shows the variation of the total l.o.s column density as a function of time for the nine l.o.s investigated in section [ 4 ] , as an example of the significant variability exhibited by the simulated wind properties along every l.o.s . not all of this variability will result in spectral variability however . lines - of - sight with total column densities that are always very compton thick ( @xmath4767@xmath32 ) are dominated by essentially nearly neutral absorption and compton scattering , irrespective of the level of variability exhibited by the simulated wind , and thus show little in the way of spectral variability . similarly , lines - of - sight that always have small total column densities ( @xmath4810@xmath49 @xmath20 , @xmath4845@xmath32 ) are too highly ionized to imprint significant features on the x - ray spectrum , again , irrespective of the level of variability exhibited by the simulated wind , and thus show little in the way of spectral variability . lines - of - sight in - between these regimes will show some significant spectral variability , with the details of the observed spectrum depending on the detailed properties of the wind material along the l.o.s at any given time . we select a single l.o.s through the transition region ( @xmath30=62@xmath32 ) and simulate the uv / x - ray spectrum for each of the first 116 snapshots from the pk04 simulation . figure [ xscorttvar ] shows the model spectra from three of the snapshots ( 800 , 844 , 895 ) chosen to highlight the dramatic spectral variability driven by the variable nature of the outflow . the spectra shown in figure [ xscorttvar ] are separated by @xmath310@xmath50s , however considerable spectral variability is observed on all timescales ; minor variability is even observed at the limit of the simulations time resolution ( 10@xmath9s ) . the larger - scale changes in spectral shape highlighted in figure [ xscorttvar ] , are driven by changes in the total column density in addition to changes in both the space density and gas temperature . over the full 10@xmath50s , the column density along the chosen l.o.s varies by over a factor 10 as shown in the top panel of figure [ tvarplot ] . this is due to changes in the density distribution along the l.o.s , which we show as a column density weighted average in the upper middle panel , while the lower middle and lower panels shows the correlated changes in ( column density weighted ) average ionization parameter and temperature , respectively . the most rapid , large - scale , variability occurs over @xmath38@xmath1010@xmath9 s ( snapshots 818 - 826 , figure [ xscortstvar ] ) , which sees a factor two change in total column density ( @xmath51=3 - 6@xmath52 @xmath20 ) drive a complete change in the character of the x - ray spectrum . such a dramatic change in the character of the x - ray spectrum has been observed in a few agn ; most notably ngc 1365 ( risilati 2007 ) , which transitioned from compton thin to compton thick ( reflection dominated ) on a timescale of 10@xmath9s , during an observation in 2006 . risilati interpret this as an occultation event occurring in an absorber close to the central engine ( @xmath310@xmath53 cm ) , a location which lends itself to an origin in a wind from the accretion disk . while qualitatively similar to the variability observed here , it is far from clear that events of this type really constitute strong evidence for the existence of radiatively driven accretion disk winds . we hope to present a more detailed discussion of the character of the spectral variability resulting from changes in the accretion disk wind in future work which , it is hoped , will help illuminate the nature and origin of events such as that observed in ngc 1365 . while the black hole mass , and the accretion rate , implicit in the hydrodynamic simulation used here makes the calculated model spectra most applicable to high accretion rate quasars ( _ e.g. _ pg1211 + 143 ) , the physical properties required to generate the spectral characteristics discussed in the following subsections may well be present in accretion disk winds from type-1 agn with a wide range of basic characteristics . interestingly , despite the impact of compton scattering on the higher inclination spectra , many of the spectra presented here display spectral features that are similar to those observed in the x - ray spectra of a range of type-1 agn . despite the differences between the black hole masses and accretion rates of these sources and that used in the pk04 simulation , here we make a qualitative comparison between the features present in the model spectra and those features that have previously been attributed to accretion disk winds in the observed sources . we hope to explore the details of the x - ray spectral features generated by winds from agn with a range of intrinsic properties in future work . gierliski & done ( 2004 ) suggested that the soft x - ray excess , observed in many ( @xmath350% - turner & pounds 1989 ) type-1 agn , may provide indirect evidence for accretion disk outflows . gierliski & done suggest absorption in a partially ionized medium with a large velocity dispersion as a possible origin for the smooth upturn in the x - ray spectrum at @xmath31 kev , and suggested that a wind from the accretion disk represents the most plausible physical origin for such material . detailed simulations presented in sd07 demonstrate that simplistic models of radiatively driven accretion disk winds , with typical physical parameters , do not have sufficient velocity dispersion to reproduce the smooth upturn of the soft x - ray excess . furthermore , sd08 showed that in order to reproduce the shape of the smooth soft x - ray excess , such an outflow must accelerate to terminal velocities of the order of @xmath30.9c , in order to produce sufficient velocity dispersion to smear out characteristic atomic features in the uv / x - ray ( again , we note that these calculations treat only radial outflows , neglecting the impact of rotation which , for significantly extended x - ray sources , may provide an additional source of velocity dispersion ) . the sd07/08 simulations thus argue against an origin in a radiatively driven or thermally driven outflow . the spectra presented in figures [ xscortlos ] & [ xscorttvar ] clearly demonstrate that physical characteristics of currently simulated accretion disk winds can not reproduce the observed properties of the soft x - ray excess , categorically ruling out such winds as the origin of this important spectral characteristic _ provided that the x - ray source is point - like with respect to ( w.r.t ) the wind material_. a point - like x - ray source is liable to be a good assumption ( for the reasons discussed in section [ 3 ] ) , however , if the x - ray source were considerably extended w.r.t the wind material , the observed spectrum would then be composed of many l.o.s that each contribute to the spectral shape . each of these l.o.s will see different regions of the outflow and , crucially , sample different regions of the outflow velocity field ( which must now include the rotational velocity components ) . the increased velocity dispersion from including the rotational velocity components will further broaden the absorption features in each l.o.s spectrum and the combination of many l.o.s will help smooth out any sharp features in the individual spectra . the resulting total spectrum would thus be expected to show both smoother and broader absorption that may well be able to reproduce the smooth shape of the soft x - ray excess . pk04 also show that larger mass black holes @xmath210@xmath54 may be able to produce considerably faster winds than the wind considered here , thus providing the necessary velocity dispersion required to smear the characteristically sharp atomic features into a smooth pseudo - continuum . the large fraction ( @xmath350% - turner & pounds 1989 ) of type-1 agn whose x - ray spectra show soft excesses presents a significant problem for both these solutions . in both cases , spectra with strong soft excesses will still only be observed from inclination angles in which the l.o.s is dominated by material in the transition region of the wind . for lower inclination l.o.s , the high ionization state and low column density of the wind material , will prevent atomic features significant enough to reproduce the soft excess from being imprinted on the spectrum , irrespective of the velocity dispersion and/or partial covering . at high inclinations , the very compton thick column density and low ionization state of the material will ensure that the observed spectrum is dominated by the emitted+scattered spectrum rather than the absorbed , transmitted , spectrum , again , irrespective of the velocity dispersion and/or partial covering . if we assume that all agn produce a line driven wind , and have x - ray emitting regions that are considerably extended w.r.t the wind material , this then predicts that , at most , @xmath313% of agn should show evidence for a strong soft x - ray excess associated with a line driven accretion disk winds . studies of galaxies in the nearby universe suggest that between 30 - 50% of agn can be classified as type-1 agn ( see , for example tajer ( 2007 ) and tozzi 2006 ) , which , coupled with the results from turner & pounds ( 1989 ) , suggest that @xmath315 - 25% of all agn in the nearby universe show evidence for a strong soft x - ray excess . furthermore , the commonality of the soft x - ray excess feature amongst nls1s ( classically thought to harbor small super - massive black holes ) , categorically rules out an origin solely in the faster line driven winds from black holes @xmath210@xmath54 . we note that turner ( 2007 ) have shown that a model based on multiple stationary layers of ionized absorbing gas can explain both the spectral shape and ( via a changing covering fraction ) the spectral variability observed in the soft x - ray band in several nearby agn ( at least one of which shows a significant soft x - ray excess in the data they use ) . however it is far from clear that the partial covering absorbers can be linked directly with the outflowing absorption responsible for the hi fe absorption lines detected in the same spectra , because the simple stationary ionized absorber models do not reproduce these features self - consistently . it remains to be seen whether a realistic model of partial covering in an accretion disk wind with an extended x - ray source is capable of reproducing the soft excess in a wide range of sources . furthermore , we note that the potential of magnetically driven wind to reproduce the soft x - ray excess through absorption remains difficult to quantify ( see sd08 ) and as such this also remains as a possible origin for this mysterious feature . the detailed properties of the high ionization ( h- & he - like - hereafter hi ) fe absorption lines present in the simulated spectra are both time , and l.o.s , dependent . figure [ feregion ] shows a close up of the hi fe absorption line features present in spectra from several l.o.s through the wind simulated in pk04 ( snapshot 800 ) . the spectral snapshots shown were specifically chosen to highlight that relatively low equivalent width ( eqw ) , hi fe absorption lines are imprinted on the transmitted spectra from a wide range of low inclination l.o.s . although the calculated model spectra most applicable to high accretion rate quasars , the prevalence of high ionization fe lines from a wide range of viewing angles in the current simulation suggests that similar physical conditions may well exist in winds from agn with considerably different masses and/or accretion rates , albeit at different inclination angle . such features may , thus , be relatively commonly observed signatures of line driven accretion disk outflows . high ionization , blue - shifted , fe absorption lines have been observed in several agn , most notably pg1211 + 143 ( pounds 2003 ) , ngc 1365 ( risaliti 2005 ) , pg1115 + 080 ( chartas 2003 ) , apm 08279 + 5255 ( chartas 2002 ) & ngc 3516 ( turner 2008 ) . four of these sources are high accretion rate quasars with black holes masses and accretion rates similar to those intrinsic to the hydrodynamical simulation used here , however ngc1365 is a classical seyfert 1 objects with a considerably smaller accretion rate ( table [ bhmasses ] ) . if interpreted as signatures of an accretion disk wind , the presence of blue - shifted , highly ionized , fe absorption lines in relatively low accretion rate sources suggests that , not only does the accretion disk in these sources generate a significant wind , albeit with a considerably slower outflow velocity , but that the physical conditions in some regions of these winds are similar ( with the exception of the velocity field ) . this is likely to be the case even for sources in which either the black hole mass , or the accretion rate , are too low to allow a ` true ' wind to be generated . in these sources , a considerable amount of material is driven to large scale - heights above the disk but the velocity of the material is insufficient for it to escape the system . despite returning to the disk at larger radii , while it is above the disk it can nevertheless imprint spectral features on the observed x - ray spectrum ( see , for example , proga 2005 ) . although the location of the region(s ) responsible for the hi fe absorption features is likely to lie at very different inclination angles for each of the sources examined here , we nevertheless make a quantitative comparison between the features observed in these agn and the calculated model spectra from our single simulation , in order to illuminate the link between the details of the observed features and their possible origin in an accretion disk wind . again , the authors note that they hope to explore the x - ray spectral features generated by winds from agn with a wide range of black hole masses and accretion rates , in detail , in future work . in the context of the pk04 wind , the relatively small eqw of the @xmath37 and @xmath38 kev lines detected in these sources , combined with the lack of strong absorption features at lower energies , suggests that the sources are viewed through the hot , low - density , flow ( which , for the simulation parameters used here lies at @xmath6 ) . this conclusion is supported by the ` doublet ' of absorption lines detected at 6.7 & 7.0 kev in both ngc 1365 and ngc 3516 , both of which have peak energies and eqws ( ngc 1365 : e@xmath55=6.697 & 6.966 kev , eqw=135 & 111 ev - risaliti 2005 . ngc 3516 : 6.700 & 6.990 kev , eqw=60 & 112 ev - turner 2008 ) similar to those of the line - like features predicted by the lower - inclination spectra presented in figure [ feregion ] ( _ e.g. _ e=6.76 & 7.05 kev and eqw=45 & 65 ev , respectively , for the 53@xmath32 l.o.s shown ) . in contrast with ngc 1365 and ngc 3516 , although a @xmath37 kev absorption feature is detected in pg1211 + 143 ( pounds 2003 , 2006 ) no corresponding 6.7 kev feature is detected in the x - ray spectrum ; a feature which is ubiquitously present in all the wind spectra simulated here . the absence of this feature may be indicative of a more complex , or a faster , outflow in pg1211 + 143 . interestingly , while the presence of the 6.7 and 7 kev lines in ngc 1365 and ngc 3516 does not rule out a l.o.s corresponding to the transition region , the properties of this region are strongly time - dependent and we would expect the observed line eqws to be variable on 10@xmath56s timescales . variability of the hi fe line profiles has already been observed in pg1211 + 143 on 10@xmath0s timescales ( reeves 2008 ) , however it remains to be seen if shorter timescale variability of the features attributed to a wind is present in the x - ray data for pg1211 + 143 , ngc 1365 of ngc 3516 . future observations of these sources hold the potential to test this prediction and , if such variability is detected , would strongly support the idea that these individual features originate in a line driven disk wind . several of the sources show narrow high eqw hi fe absorption lines that are very strongly blue - shifted , most notably pg1115 + 080 ( [email protected] kev , [email protected] kev ) and apm 08279 + 5255 ( [email protected] kev , [email protected] kev ) . while the model spectra from the pk04 simulation do not produce such strong , highly blue - shifted lines , it is plausible that future simulations of a faster wind , originating from a larger mass black hole , could well produce the @xmath39 kev features . it currently remains unclear , however , whether it is plausible to produce a fast _ enough _ wind through radiation driving , or if the line - like features in the resulting spectrum would remain narrow , given the strong acceleration gradient that must be present in such an outflow . the most rapid and , arguably , the most powerful outflow identified from absorption features in the uv / x - ray spectra of nearby agn is that of pds 456 ( reeves 2003 , obrien 2008 ) . reeves attribute the deep 7 - 13 kev absorption trough observed in the x - ray spectrum to a series of blue - shifted k - shell absorption edges from hi fe . the form of the @xmath27 kev spectrum appears to be extremely similar to the simulated spectra from l.o.s passing through the transition region , presented both in figure [ xscortlos ] ( 62@xmath32 & 65@xmath32 ) and figure [ xscorttvar ] ( snapshot 844 & 895 ) , demonstrating that the 50,000 km s@xmath4 outflow velocity suggested in reeves ( 2003 ) is not required to make this feature . however the model spectra presented in figure [ xscortlos ] & [ xscorttvar ] all show strong , relatively narrow , absorption features between @xmath36 - 7 kev which are notably absent in the observed x - ray spectrum of pds 456 . the ubiquity of these narrow lines in all the wind spectra calculated from the pk04 simulation , drives us to the conclusion that the high energy features observed in pds 456 can not be produced by the line driven disk wind parameters explored in this simulation . an origin in a considerably faster wind ( with a velocity more similar to the 50,000 km s@xmath4 originally suggested by reeves ) remains a plausible explanation , however , whether it is possible to produce such a wind through line driving is unclear . in order to shift the 6 - 7 kev features to the required energies , such a wind would require a typical outflow velocity more than five times that present in the pk04 simulations . furthermore , if the high energy spectrum of pds 456 does have an origin in the transition region of an accretion disk wind , we would expect it to show considerable variability . once again , future observations of pds 456 hold the potential to test this prediction and , if such variability is detected , would strongly support the idea that this features originates in a line driven disk wind . we note that while it is possible for to resolve the 6.7/7 kev line present in the lower - inclination spectra ( figure [ feregion ] ) , the details of the more complex iron absorption structures seen in the higher inclination spectra can not be resolved with data from current x - ray satellite data . broad , blue - shifted , absorption lines detected in the optical / uv spectra of broad absorption line ( bal ) qsos are strong evidence for accretion disk outflows ( _ e.g. _ weymann 1991 , turnshek 1996 , etc ) . previous accretion disk wind models have been able to match some of the wind properties inferred from individual models to the properties of detected spectral features ( _ e.g. _ murray 1995 , konigl & kartje 1994 , pk04 etc ) , however none of these models predict the full optical / uv spectrum seen through a self - consistent wind and , thus , the true origin of these spectral features remains somewhat ambiguous . bal quasar features have been linked to line driven winds in particular , based on the ability of the numerical simulations to produce a self - shielded wind and a fast , high column , moderately ionized flow with a relatively small angular size ( _ i.e. _ the transition region - pk04 , proga 2005 , etc ) . in addition , the properties of the bals in a large sample of sloan digital sky survey quasars have been shown to be strongly correlated with the uv continuum properties , suggesting an origin in a uv - driven wind ( ganguly 2007 ) . the spectra calculated here , using the self - consistent wind properties from the pk04 numerical simulations , do not show any strong absorption features at wavelengths @xmath2300 , except for l.o.s @xmath267@xmath32 ; where the column densities are sufficiently large as to ensure that the uv spectra are dominated by scattered+emitted flux . this result runs contrary to the broad uv line profiles , presented in pk04 , predicted by the same simulations . the cause of this discrepancy lies in the way treats the uv continuum emission . for the spectra calculated here , we assume that the uv continuum is simply an extension of the x - ray power - law . while the resulting spectra are thus a reasonable representation of the l.o.s transmitted x - ray spectrum , the same can not be said for the simulated uv spectrum , because the majority of the uv flux from agn actually originates from an extended region of the accretion disk itself . currently , neglects the uv photons that originate from the accretion disk , primarily because the code is not optimized to deal effectively with an extended continuum source . the extended nature of the uv emission from the disk would require not only the calculation of multiple l.o.s , but would also require that the full 3-d velocity field be accounted for in calculating the spectral contribution from each part of the extended source . for this reason , we do not discuss the uv spectrum predicted by in detail . we hope to address the issues relating to both an extended x - ray source ( see section [ 3 ] ) , and the extended uv emission from the accretion disk in future version of the code . evidence of x - ray warm absorption has been found in about 50% of nearby seyfert 1 galaxies using spectra ( reynolds 1997 ) . an origin in an accretion disk wind has often been implicated for this warm absorption ( _ e.g. _ pg 0844 + 349 & pg1211 + 143 - blustin 2005 , ngc 3783 - kaspi 2002 , ngc 4051 - krongold 2007 ) however a careful examination of the properties of a large sample of nearby agn suggests that , in most cases , these features are more likely to originate from outflow on larger scales ( blustin 2005 ) . in both cases , however , the analysis of these features typically relies on photoionization codes that are not optimized for calculating the correct spectral shape associated with a rapidly accelerating outflow , whatever its origin . such codes typically assume a single outflow velocity and a uniform medium when calculating a model spectrum . several of these models are then used to reproduce the complex observed x - ray spectral shape , resulting in a ` many - phase ' outflow with kinematically and energetically separated components attributed to a wind . provided the densities of the regions are known , their distance from the central ionizing continuum source can then be calculated from the best fit ionization states and the source luminosity . neglecting the acceleration ( i.e. , the velocity shear ) in the wind , as well as assuming multiple , decoupled , media with uniform densities and ionization states , clearly represents a poor way of modeling the spectral signatures imprinted by a wind ( an in particular an accretion disk wind , where the velocity shear is large ) . interestingly , some of the accretion disk wind spectra presented here do show features that are qualitatively similar to typical observed warm absorber however the features are limited to a small range of viewing angles . specifically , the dominance of compton thick absorption in the dense , warm , equatorial flow ( @xmath30@xmath267@xmath57 ) , and the lack of spectral features imprinted by the hot , rarefied , polar flow ( @xmath30@xmath5855@xmath57 ) , limits the presence of warm absorber features to the transition region in - between . clearly the 50% rate of warm absorbers in type-1 agn can not be accounted for by the limited inclination angles which pass through the transition region ruling out an origin in a radiatively driven wind for most warm absorbers . it remains plausible , however , that a small number of agn may have warm absorber features imprinted on their x - ray spectra by an accretion disk wind . unfortunately , the massive spectral variability generated by changes in the composition of the transition region ( see figure [ xscorttvar ] ) implies that any warm absorber features that may originate in this region would be considerably variable on 10@xmath9s timescales and , in some cases will vanish altogether , placing yet further constrains on the likelihood of observing warm absorber features associated with an accretion disk wind . finally , detailed hydrodynamic simulations of outflows from the surface of the dusty molecular torus which have shown that this material has the correct physical and kinematic properties required to reproduce typical warm absorber features ( dorodnitsyn , kallman & proga 2008a , 2008b ) . unfortunately , only tentative spectra are presented in dorodnitsyn , kallman & proga ( 2008b ) and the authors have not presented ( yet ) a detailed description of how the spectra are calculated or how well their characteristics match that of observed warm absorber . while details of the l.o.s spectra generated from these simulations are not yet fully available , the combined weight of evidence seems to favour an origin for the warm absorber features in an outflow from the torus . we caution , however , that this debate is far from settled . everett ( 2005 ) , for example , present a dynamical model of the warm absorber based on magneto - centrifugal winds . although this model appears to be reasonable , we do not know what spectra this model predicts and it therefore remains unclear whether this model does in fact explain the observations . we present spectra from a series of simulations that use the numerical accretion disk wind simulation presented in pk04 to provide a set of self - consistent physical properties for the outflow . this addresses many of the problems inherent to previous simulations in which the density structure , velocity field , total column and geometrical thickness were all chosen in an ad - hoc manner ( sd07 , sd08 ) . while the black hole mass , and the accretion rate , implicit in the hydrodynamic simulation used here , makes the calculated model spectra most applicable to high accretion rate quasars ( _ e.g. _ pg1211 + 143 ) , the physical properties required to generate the observed spectral characteristics may well be present in accretion disk winds from agn with differing basic characteristics , and thus may be observed in a wide range of type-1 agn ( and this is indeed the case , see section 6 ) . the model spectra presented here are calculated for a range of lines - of - sight though the hydrodynamically simulated wind ( figure [ xscortlos ] ) and for a range of simulation snapshots ( figure [ xscorttvar ] ) . the properties of these spectra clearly reflect the three distinct regions apparent in the pk04 simulations : \(1 ) simulated x - ray spectra from lines - of - sight through the hot , low - density , polar flow ( @xmath6 ) remain largely featureless , despite the more equatorial lines - of - sight having column densities as large as @xmath31.5@xmath1010@xmath33 @xmath20 . this is due primarily to the low physical densities ( n@xmath3410@xmath0 @xmath12 ) and high temperatures ( t@xmath4010@xmath59 k ) of the gas in this region , which result in an extremely high ionization state throughout the material . despite being largely featureless , particularly in the uv band , spectra from this region do show highly - ionized , blue - shifted , fe absorption features in the 6.7 - 9 kev range which are qualitatively similar to those detected in the x - ray spectra of a growing number of agn . \(2 ) the model spectra from lines - of - sight through the dense , warm , fast equatorial flow ( @xmath7 ) have column densities of @xmath24@xmath1010@xmath36 @xmath20 which results in almost complete attenuation of the spectrum by compton scattering . the majority of the material along the l.o.s is essentially neutral ( log@xmath39(@xmath24)@xmath3 - 3 ) , resulting in considerable absorption being imprinted on the x - ray spectrum . for such a l.o.s , the scattered+emitted flux is expected to dominate the uv /x - ray spectrum ( see sim 2008 for a discussion of the properties of the scattered+emitted x - ray spectrum from a variety of idealized accretion disk winds ) . if such outflows are common in agn , the opening angle of this flow would imply that @xmath325% of agn should be very compton thick , even without a contribution from the putative molecular torus . \(3 ) simulated x - ray spectra from lines - of - sight through the transition region ( @xmath8 ) , that separates the two larger regions of the simulated outflow , result in spectra which show considerable atomic features across the 0.3 - 13 kev range . the physical properties of the material in this region are broadly similar to those found in the hot , low - density , polar flow ( t@xmath4010@xmath41 k , [email protected] ) except that the l.o.s gas density distribution has both a higher maximum ( n@xmath4310@xmath44 @xmath12 ) and , crucially , significantly more high - density cells , than l.o.s through the polar regions . the significantly greater amount of material along the l.o.s ( column densities of [email protected]@xmath1010@xmath46@xmath20 ) means that the majority of this region has an intermediate ionization parameter ( log@xmath39(@xmath24)@xmath33.5 - 4.5 ) and imprints a wide range of atomic features on the spectrum . the spectra from this region also show highly - ionized , blue - shifted , fe absorption features in the 6.7 - 9 kev range however their properties are markedly different from the features associated with the hot , low - density , polar flow . importantly , the spectra presented here clearly demonstrate that current models of line driven agn accretion disk winds do not have the appropriate physical characteristics required to reproduce the smooth soft x - ray excess , assuming a point - like x - ray source . for the case of a considerably extended emission region , the inclusion of rotational velocity component , coupled with the effects of partial covering , will change the spectrum we observe . we note that turner ( 2007 ) have shown that a model based on multiple stationary layers of ionized absorbing gas can explain both the spectral shape and ( via a changing covering fraction ) the spectral variability observed in several nearby agn ( at least one of which shows a significant soft x - ray excess in the data they use ) . however it remains to be seen whether a realistic model of partial covering in an accretion disk wind with an extended x - ray source is capable of reproducing the soft excess in a wide range of sources . testing this idea is something we aim to address with future version of the code . some of the spectra simulated here do show features similar to those observed in the x - ray spectra of some quasars , particularly the presence of blue - shifted , high ionization , fe lines . in particular , the line centroid energies and equivalent widths are quantitatively similar to lines observed in some nearby quasars ( _ e.g. _ pg1211 + 143 - pounds 2003 & ngc 1365 - risaliti 2005 ) . the models do not reproduce the @xmath28 kev lines observed in some high redshift quasars ( _ e.g. _ pg1115 + 080 - chartas 2003 & apm 08279 + 5255 - chartas 2002 ) or the deep @xmath37 - 13 kev absorption trough observed in the x - ray spectrum of pds 456 ( reeves 2003 ) , however a considerably faster wind than the one investigated here may well be able to reproduce these features . interestingly , the l.o.s outflow properties through the transition region show considerable variability on all the timescales testable with the pk04 simulation data ( figures [ xscorttvar ] & [ xscortstvar ] ) , suggesting that features associated with this region , in particular the high ionization fe absorption lines , should show corresponding spectral variability . contrary to previous models , the simulated spectra do not show any strong absorption features at uv wavelengths for spectra that are not dominated by compton scattering ( where the scattered+emitted flux will dominate the uv / x - ray spectrum ) , primarily due to the assumed form and physical origin of the uv continuum within ( see section [ 6 - 1 ] ) . as such , it remains plausible that uv bals could originate from a line driven disk wind in agn , as shown in pk04 ( see figure . 3 therein ) . the compton think winds simulated here predict that high accretion rate ( l / l@xmath1 ) agn are likely to be strongly effected by obscuration , in sharp contrast to the cleaner picture that is generally assumed based on the observed relation between the opening angle of the molecular torus and the agn luminosity ( see polletta 2008 , and references therein ) . the authors thank tim kallman for writing and supporting the code . njs acknowledges financial support through a uk - china fellowship for excellence and cd acknowledges financial support through an stfc senior fellowship . dp acknowledges support provided by the national aeronautics and space administration through the chandra award tm7 - 8008x issued by the chandra x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 39073 . this research has made extensive use of nasa s astrophysics data system abstract service . blustin a. j. , page m. j. , fuerst s. v. , branduardi - raymont g. & ashton c. e. , 2005 , a&a , 431 , 111 chartas g. , brandt w. n. & gallagher s. c , 2003 , apj , 595 , 85 chartas g. , brandt w. n. , gallagher s. c & garmire g. p. , 2002 , apj , 579 , 169 dorodnitsyn a. , kallman t. & proga d. , 2008 , apjl , 675 , l5 ganguly r. , brotherton m. s. , cales s. , scoggins b. , shang z. & vestergaard m. , 2007 , apj , 665 , 990 gierliski m. & done c. , 2004 , mnras 349 , l7 kaspi s. , brandt w. n. , george i. m. , netzer h. , crenshaw d. m. , gabel j. r. , hamann f. w. , kaiser m. e. , koratkar a. , kraemer s. b. , kriss g. a. , mathur s. , mushotzky r. f. , nandra k. , peterson b. m. , shields j. c. , turner t. j. & zheng w. , 2002 , apj , 574 , 643 konigl a. & kartje j. f. , 1994 , apj , 434 , 446 krongold y. , nicastro f. , elvis m. , brickhouse n. , binette l. , mathur s. & jimnez - bailn e. , 2007 , apj , 659 , 1022 murray n. , chiang j. , grossman s. a. & voit g. m. , 1995 , apj , 451 , 498 nikoajuk m. , czerny b. , zikowski j. & gierliski m. , 2006 , mnras , 370 , 1534 obrien p. t. , reeves j. n. , simpson c. & ward m. j. , 2005 , mnras , 360 , l25 pounds k. a. & page k. l. , 2006 , mnras , 372 , 1275 pounds k. a. , reeves j. n. , king a. r. , page k. l. , obrien p. t. & turner m. j. l. , 2003 , 345 , 705 proga d. , 2005 , apjl , 630 , l9 proga d. & kallman t. r. , 2004 , apj , 616 , 688 proga d. , stone j. m. & kallman t. r. , 2000 , apj , 543 , 686 reeves j. , done c. , pounds k. , terashima y. , hayashida k. , anabuki n. , uchino m. & turner m. , 2008 , mnras , 385 , l108 reeves j. n. , obrien p. t. & ward m. j. , 2003 , apjl , 593 , l65 risaliti g. , bianchi s. , matt g. , baldi a. , elvis m. , fabbiano g. & zezas a. , 2005 , apjl , 630 , l129 risaliti g. , elvis m. , fabbiano g. , baldi a. , zezas a. & salvati m. , 2007 , apj , 659 , l111 schurch n. j. & done c. , 2007 , mnras , 381 , 1413 - sd07 schurch n. j. & done c. , 2008 , mnras , 386 , l1 - sd08 tajer m. , polletta m. , chiappetti l. , maraschi l. , trinchieri g. , maccagni d. , andreon s. , garcet o. , surdej j. , pierre m. , le fvre o. , franceschini a. , lonsdale c. j. , surace j. a. , shupe d. l. , fang f. , rowan - robinson m. , smith h. e. & tresse l. , 2007 , a&a , 467 , 73 turner t. j. , miller l. , reeves j. n. & kraemer s. b. , 2007 , a&a , 475 , 121 turner t. j. & pounds k. a. , 1989 , mnras , 240 , 833 turner t. j. , reeves j. n. , kraemer s. b. & miller l. , 2008 , a&a , 483 , 161 turnshek d. a. , kopko m. j. , monier e. , noll d. , espey b. r. & weymann r. j. , 1996 , apj , 463 , 110 weymann r. j. , morris s. l. , foltz c. b. & hewett p. c. , 1991 , apj , 373 , 23 ) from five different lines - of - sight , with different inclinations , through the hydrodynamic simulation of an agn outflow presented in pk04 . the spectra were calculated with ( v5.18 ) using snapshot 800 of the pk04 simulation to provide self - consistent physical outflow properties . the input ionizing power - law spectrum is shown in black . the spectra correspond to inclinations of @xmath30 = 50@xmath32 ( red ) , 57@xmath32 ( green ) , 62@xmath32 ( blue ) , 65@xmath32 ( cyan ) & 67@xmath32 ( purple ) . these l.o.s were chosen to highlight the range of spectral shapes that result from the different physical properties throughout the simulated wind.,height=15 ] ( 30@xmath32 - black , 50@xmath32 - red , 57@xmath32 - green , 61@xmath32 - dark blue , 62@xmath32 - light blue [ see figure [ tvarplot ] , _ top panel _ ] , 64@xmath32 - pink , 65@xmath32 - yellow , 67@xmath32 - orange , 75@xmath32 - purple ) . column densities @xmath4710@xmath36 @xmath20 are extremely compton thick and the resultant spectrum is always dominated by essentially nearly neutral absorption and compton scattering . column densities @xmath4810@xmath49 @xmath20 are too highly ionized to imprint significant features on the x - ray spectrum . columns densities in - between will imprint significant features on the x - ray spectrum , ranging from a few isolated line feature ( in particular high ionization fe absorption lines ) to significant attenuation by an almost neutral column.,width=9 ] ) from a 62@xmath32 line - of - sight through the initial t=0 snapshot ( 800 , red ) , the t=4.4@xmath1010@xmath60s snapshot ( 844 , green ) and the t=9.5@xmath1010@xmath60s snapshot ( 895 , blue ) of the hydrodynamic simulation presented in pk04 . the spectra were calculated with ( v5.18 ) using snapshots of the pk04 simulation to provide self - consistent physical outflow properties . the input ionizing power - law spectrum is shown in black . both the l.o.s and the individual snapshots were chosen to highlight the wide range of spectral shape variability that results from the time - variable physical properties of the simulated wind along the l.o.s.,height=15 ] l.o.s through the pk04 simulation . _ top panel _ : the total l.o.s column density . _ upper - middle panel _ : the column - density - weighted average space density ( @xmath58n@xmath61 ) . _ lower - middle panel _ : the column - density - weighted average ionization parameter ( @xmath58log@xmath62@xmath61 ) . _ bottom panel _ : the column - density - weighted average temperature ( @xmath58t@xmath61 ) . significant changes in the character of the simulated x - ray spectrum are driven by the variability of these properties . significant changes in the total column density and the column - density - weighted average space density , which in turn drive changes in the column - density - weighted average ionization parameter , are of particular importance.,width=17 ] ) from a 62@xmath32 line - of - sight through the t=1.8@xmath1010@xmath60s snapshot ( 818 , red ) , the t=2.6@xmath1010@xmath60s snapshot ( 826 , green ) of the hydrodynamic simulation presented in pk04 . the spectra were calculated with ( v5.18 ) using snapshots of the pk04 simulation to provide self - consistent physical outflow properties . the input ionizing power - law spectrum is shown in black . these snapshots were chosen to highlight the dramatic short timescale ( @xmath310@xmath60s ) spectral variability that results from the short time variability of the simulated wind.,height=15 ] ) from four different lines - of - sight , with different inclinations , through the hydrodynamic simulation of an agn outflow presented in pk04 . the spectra were calculated with ( v5.18 ) using snapshot 800 of the pk04 simulation to provide self - consistent physical outflow properties . the spectra correspond to inclinations of @xmath30 = 50@xmath32 ( red ) , 53@xmath32 ( green ) , 57@xmath32 ( blue ) & 62@xmath32 ( cyan ) . these l.o.s were chosen to highlight the details of high ionization fe absorption lines that are produced by the simulated wind.,height=15 ]

we use , together with the hydrodynamic accretion disc wind simulation from proga & kallman ( 2004 ) , to calculate the impact that the accretion disk wind has on the x - ray spectrum from a 10@xmath0 solar mass black hole active galactic nuclei ( agn ) accreting at 0.5 l / l@xmath1 . the numerical simulation provides a set of self - consistent physical properties for the outflow that mitigates many of the problems inherent to previous simulations . the properties of the resulting spectra depend on viewing angle and clearly reflect the distinct regions apparent in the original hydrodynamic simulation . very equatorial lines - of - sight ( l.o.s ) encounter extremely compton thick column densities and produce spectra that are dominated by compton scattering and nearly - neutral absorption . polar l.o.s encounter small , highly ionized , column densities , and result in largely featureless spectra . finally , l.o.s that intersect the transition region between these extremes encounter moderately ionized , marginally compton thick column densities , that imprint a wide range of absorption features on the spectrum . both polar and transition region l.o.s produce spectra that show highly - ionized , blue - shifted , fe absorption features that are qualitatively similar to features observed in the x - ray spectra of a growing number of agn . the spectra can not reproduce the @xmath28 kev lines , or the deep @xmath37 - 13 kev absorption trough , observed in some high redshift quasars , although a considerably faster wind may well be able to reproduce these features . the spectra presented here clearly demonstrate that current simulations of line driven agn accretion disk winds can not reproduce the smooth soft x - ray excess . furthermore , they predict that high accretion rate ( l / l@xmath1 ) agn are likely to be strongly affected by obscuration , in sharp contrast to the clean picture that is generally assumed , based on the observed relation between the opening angle of the molecular torus and agn luminosity .

a group is _ coherent _ if all its finitely generated subgroups are finitely presented . the best known examples of coherent groups are free groups , surface groups , polycyclic groups , and @xmath1-manifold groups . outside of these examples , few criteria for determining the coherence or incoherence of an arbitrary group presentation are known . free groups are easily proven to be coherent by observing that subgroups of free groups are free and hence finitely presented if they are finitely generated . similarly , surface groups and polycyclic groups are easily shown to be coherent . the coherence of the fundamental groups of @xmath1-manifolds is a deeper result proved independently by scott @xcite and shalen ( unpublished ) . theoretical interest in the coherence of various groups has been prompted in part by a desire to perform calculations . groups in which all of the finitely generated subgroups have a computable finite presentation are especially amenable to computer investigation . the range of possible positive results is limited by the existence of various counterexamples . rips has produced examples of incoherent word - hyperbolic groups @xcite , bestvina and brady have produced examples of incoherent right - angled artin groups @xcite , and wise has produced examples of compact negatively curved @xmath0-complexes with incoherent fundamental groups @xcite . recently , feighn and handel proved the remarkable positive result that the mapping torus of any injective endomorphism of a free group is coherent @xcite . their theorem is related to the coherence of @xmath1-manifolds in the following sense . many @xmath1-manifolds arise as surface bundles over a circle , and their fundamental groups are thus isomorphic to extensions of surface groups by @xmath2 . the result of @xcite shows that extensions of free groups by @xmath2 are also coherent , thus extending the often successful analogy between free groups and surface groups . the current investigation was primarily motivated by the following open problem which has remained unresolved for over thirty years with very little forward progress : problem(g.baumslag , @xcite ) is every one - relator group coherent ? in this article we describe a criterion which , if successful , allows one to conclude that the group under consideration is coherent . the criterion involves a new notion which we call `` perimeter '' . roughly speaking , given a map @xmath3 between 2-complexes , the perimeter of @xmath4 which we denote by @xmath5 , is a measure of how large the `` boundary '' of @xmath4 is relative to @xmath6 . the strategy underlying the results in this paper is that if @xmath3 is unsatisfactory , because for instance , it is not @xmath7-injective , then some @xmath0-cells can be added to @xmath4 which reduce the perimeter , and after repeating this finitely many times , we obtain a satisfactory map between @xmath0-complexes . we will now state several of our main results . all of the undefined terminology , such as perimeter , weighted @xmath0-complex , and the various hypotheses will be explained in the course of the article . the main coherence result is the following : let @xmath6 be a weighted @xmath0-complex which satisfies the perimeter reduction hypothesis . \a ) if @xmath4 is a compact connected subcomplex of a cover @xmath8 of @xmath6 , and the inclusion @xmath9 is not @xmath7-injective , then @xmath4 is contained in a compact connected subcomplex @xmath10 such that @xmath11 . \b ) for any compact subcomplex @xmath12 , there exists a compact connected subcomplex @xmath4 containing @xmath13 , such that @xmath5 is minimal among all such compact connected subcomplexes containing @xmath13 . consequently @xmath14 is coherent . a geometric consequence of theorem [ thm : gct ] is the following : let @xmath6 be a weighted aspherical @xmath0-complex which satisfies the perimeter reduction hypothesis . if @xmath15 is a covering space and @xmath16 is finitely generated , then every compact subcomplex of @xmath8 is contained in a compact core of @xmath8 . a subspace @xmath4 of a geodesic metric space @xmath6 is quasiconvex if there is an @xmath17 neighborhood of @xmath4 which contains all of the geodesics in @xmath6 which start and end in @xmath4 . in group theory , a subgroup @xmath18 of a group @xmath19 generated by @xmath20 is quasiconvex if the @xmath21-cells corresponding to @xmath18 form a quasiconvex subspace of the cayley graph @xmath22 . the next main result in this article is a criterion which , if satisfied , allows one to conclude that a group is locally quasiconvex , i.e. that all finitely generated subgroups are quasiconvex . as the reader will observe from some of the applications described below , most of the groups which we can show are coherent also satisfy this stronger criterion , and thus will be locally quasiconvex as well . the exact relationship between the two criteria will become clear in the course of the article . our main quasiconvexity theorem is the following : let @xmath6 be a compact weighted @xmath0-complex . if @xmath6 satisfies the straightening hypothesis , then every finitely generated subgroup of @xmath14 embeds by a quasi - isometry . furthermore , if @xmath14 is word - hyperbolic then it is locally quasiconvex . the statements of the perimeter reduction hypothesis definition [ def : prh ] and the straightening hypothesis ( definition [ def : qrh ] ) are rather technical , but the strength of the theorems above can be illustrated easily through some explicit consequences that we now describe . first of all , as a consequence of theorem [ thm : gct ] we obtain the following result : let @xmath23 be a cyclically reduced word and let @xmath24 . if @xmath25 , then @xmath19 is coherent . in particular , for every word @xmath23 , the group @xmath24 is coherent provided that @xmath26 is sufficiently large . with a slightly stronger requirement on the degree of torsion , we can obtain the following consequence of theorem [ thm : lqc ] : let @xmath27 be a one - relator group with @xmath28 . then @xmath19 is locally quasiconvex . a similar result holds for multi - relator groups : let @xmath29 be a finite presentation , where each @xmath30 is a cyclically reduced word which is not a proper power . if @xmath30 is not freely conjugate to @xmath31 for @xmath32 , then there exists a number @xmath33 such that for all choices of integers @xmath34 the group @xmath35 is coherent . specifically , the number @xmath36 has this property , where @xmath37 and @xmath38 denote longest and shortest words among the @xmath30 , respectively . moreover , if @xmath39 for all @xmath40 , then @xmath19 is locally quasiconvex . a small - cancellation application with a different flavor is the following : let @xmath41 be a small cancellation presentation which satisfies @xmath42 . if each @xmath43 occurs at most @xmath44 times among the @xmath45 , then @xmath19 is coherent and locally quasiconvex . more precise applications to additional groups which are important in geometric group theory can be found in sections [ sec:1relator ] , [ sec:2-cell - small ] , [ sec : sc - lqc ] , and [ sec : manifold ] . many of the individual results derived in these sections can be summarized by the following qualitative description : qualitative summary if a presentation has a large number of generators relative to the sum of the lengths of the relators , and the relators are relatively long and sufficiently spread out among the generators , then the group is coherent and locally quasiconvex . the main triumph of these ideas , is that while we have only partially solved baumslag s problem , we have substantially answered the problem raised by c.t.c . wall of whether small - cancellation groups are coherent @xcite . in a separate paper @xcite , we give a much more detailed application of the strongest results in this paper to small - cancellation groups . furthermore , families of examples are constructed there which show that the applications to small - cancellation theory are asymptotically sharp . an application of our theory towards the local quasiconvexity of one - relator groups with torsion is given in @xcite , an application towards the coherence of various other one - relator groups is given in @xcite , and an application towards the subgroup separability of coxeter groups is given in @xcite . we conclude this introduction with a brief section - by - section description of the article . the concept of the perimeter of a map is introduced in section [ sec : perimeter ] and it is from this concept that all of our positive results are derived . section [ sec : gch ] shows how this concept leads to the notion of a perimeter reduction and it also contains our general coherence theorem . sections [ sec : attachments ] , [ sec:2cell ] , and [ sec : pathct ] develop the specific case where the perimeter can be reduced through the addition of a single @xmath0-cell . the proofs lead to procedures which are completely algorithmic . these algorithmic approaches are described in section [ sec : algorithms ] . in particular we show that for the groups included in the @xmath0-cell coherence theorem there exists an algorithm to compute an explicit finite presentation for an arbitrary finitely generated subgroup . additionally , we show that the time it takes to produce such a finite presentation is quadratic in the total length of its set of generators . section [ sec : pathct ] presents a more technical theorem about coherence using sequences of paths . in section [ sec:1relator ] the theory developed in the first half of the article is applied to the class of one - relator groups with torsion . similarly , section [ sec:2-cell - small ] presents some background on small cancellation theory , and gives some applications of the theory to the coherence of small cancellation groups . as this work has evolved over the past six years , two things have become clear : first of all , a much richer collection of positive results can be obtained by attaching `` fans '' of 2-cells to reduce perimeter instead of attaching single @xmath0-cells . secondly , the most significant conclusion of the theory appears to be local quasiconvexity rather than coherence , and in fact , we know of no word - hyperbolic group which satisfies our coherence criterion but which is not locally - quasiconvex as well . the latter part of the paper introduces fans in the context of coherence theorems that utilize them . thereafter , fans are employed in the statements and proofs of the local quasiconvexity theorems . fans are defined in section [ sec : fanct ] , where the general theory is extended by incorporating fans into the statements and arguments . section [ sec : quasi ] contains basic definitions and results about quasiconvexity which we will need , and section [ sec : lqc - main ] presents our main theorem about local quasiconvexity . in section [ sec : sc - lqc ] we return to small cancellation groups , and give additional coherence applications as well as some local quasiconvexity applications . section [ sec : manifold ] uses these applications to small cancellation groups to obtain results about @xmath1-manifold groups . while the coherence of @xmath1-manifold groups has long been known , when successful , our method gives a different approach towards understanding the reasons behind this remarkable theorem . finally , in section [ sec : other ] , we describe theorems and algorithms related to the finitely generated intersection property and the generalized word problem . the main goal of this section is to introduce the notion of the perimeter of a @xmath0-complex @xmath4 relative to a particular map @xmath46 . we begin with a number of basic definitions and the definition of the unit perimeter . in the second half of the section we broaden the definition to allow for the introduction of weights . the weighted perimeter of a @xmath0-complex is a measure of the complexity of a map which will be used to prove coherence and local quasiconvexity theorems throughout the article . [ def : combinatorial ] a map @xmath3 between cw complexes is _ combinatorial _ if its restriction to each open cell of @xmath4 is a homeomorphism onto an open cell of @xmath6 . a cw complex @xmath6 is _ combinatorial _ provided that the attaching map of each open cell of @xmath6 is combinatorial for a suitable subdivision . it will be convenient to be explicit about the cells in a combinatorial @xmath0-complex . a _ polygon _ is a @xmath0-dimensional disc whose cell structure has @xmath26 @xmath21-cells , @xmath26 @xmath47-cells , and one @xmath0-cell where @xmath48 is a natural number . if @xmath6 is a combinatorial @xmath0-complex then for each open @xmath0-cell @xmath49 there is a polygon @xmath50 , a combinatorial map @xmath51 and a map @xmath52 such that the diagram @xmath53 commutes , and the restriction @xmath54 is the attaching map of @xmath13 . in this article the term _ @xmath0-cell _ will always mean a combinatorial map @xmath51 where @xmath50 is a polygon . the corresponding _ open @xmath0-cell _ is the image of the interior of @xmath50 . a similar convention applies to @xmath47-cells . let @xmath55 denote the graph with two @xmath21-cells and one @xmath47-cell connecting them . since combinatorial maps from @xmath55 to @xmath6 are in one - to - one correspondence with the characteristic maps of @xmath47-cells of @xmath6 , we will often refer to a map @xmath56 as a _ @xmath47-cell _ of @xmath6 . in the study of infinite groups , the most commonly considered combinatorial @xmath0-complexes correspond to presentations . recall that the _ standard @xmath0-complex _ of a presentation is formed by taking a unique @xmath21-cell , adding a labeled oriented @xmath47-cell for each generator , and then attaching a @xmath0-cell along the closed combinatorial path corresponding to each relator . [ conv : combinatorial ] unless noted otherwise , all complexes in this article are combinatorial @xmath0-complexes , and all maps between complexes are combinatorial maps . in addition , we will avoid certain technical difficulties by always assuming that all of the attaching maps for the @xmath0-cells are immersions . for @xmath0-complexes with a unique @xmath21-cell , this is equivalent to allowing only cyclically reduced relators in the corresponding presentation . [ def : basic ] a local injection between topological spaces is an _ immersion . _ if @xmath46 is an immersion on @xmath57 , then @xmath58 is a _ near - immersion_. if @xmath59 is an immersion on @xmath60 then @xmath58 is a _ @xmath47-immersion_. let @xmath61 be the induced homomorphism between fundamental groups . the map @xmath58 is _ @xmath7-injective _ [ respectively _ @xmath7-surjective _ ] if @xmath62 is injective [ surjective ] . finally , if @xmath46 and @xmath63 are fixed maps , then a map @xmath64 is a _ lift of @xmath65 _ or _ a lift of @xmath66 to @xmath4 _ whenever the composition @xmath67 . a _ path _ is a map @xmath68 where @xmath69 is a subdivided interval or a single @xmath21-cell . in the latter case , @xmath69 is a _ trivial path_. a _ cycle _ is a map @xmath70 where @xmath13 is a subdivided circle . given two paths @xmath71 and @xmath72 such that the terminal point of @xmath69 and the initial point of @xmath73 map to the same @xmath21-cell of @xmath6 , their concatenation @xmath74 is the obvious path whose domain is the union of @xmath69 and @xmath73 along these points . the path @xmath68 is a _ closed path _ provided that the endpoints of @xmath69 map to the same @xmath21-cell of @xmath6 . a path or cycle is _ simple _ if the map is injective on @xmath21-cells . the _ length _ of the path @xmath69 or cycle @xmath13 is the number of @xmath47-cells in the domain and it is denoted by @xmath75 or @xmath76 . the _ interior _ of a path is the path minus its endpoints . in particular , the @xmath21-cells in the interior of a path are the @xmath21-cells other than the endpoints . a _ subpath _ @xmath73 of a path @xmath69 [ or a cycle @xmath13 ] is given by a path @xmath78 [ @xmath79 in which distinct @xmath47-cells of @xmath73 are sent to distinct @xmath47-cells of @xmath69 [ @xmath13 ] . notice that the length of a subpath is at most that of the path [ cycle ] which contains it . finally , note that any nontrivial closed path determines a cycle in the obvious way . finally , when the target space is understood we will often just refer to @xmath68 as the path @xmath69 . the letters @xmath6 and @xmath4 will always refer to spaces , @xmath50 will always refer to a closed @xmath0-cell , and @xmath69 will always denote a path . we follow the convention that lowercase letters ( such as @xmath80 , @xmath81 , and @xmath82 ) refer to specified @xmath47-cells in the space denoted by the corresponding uppercase letter . thus @xmath82 is a @xmath47-cell in the boundary of the @xmath0-cell @xmath50 and @xmath80 is a @xmath47-cell in the space @xmath6 . we will be very interested in examining the behavior of maps and spaces along selected @xmath47-cells . accordingly , the pair @xmath83 will denote a space together with a chosen @xmath47-cell in that space , and we will write @xmath84 to denote a map @xmath85 with the property that @xmath86 . [ def : sides ] let @xmath6 be a fixed @xmath0-complex , and let @xmath50 be a @xmath0-cell of @xmath6 . let @xmath82 be a @xmath47-cell in @xmath87 and let @xmath80 be the image of @xmath82 in @xmath6 . the pair @xmath88 will then be called a _ side of a @xmath0-cell of @xmath6 which is present at @xmath80_. the collection of all sides of @xmath6 which are present at @xmath80 will be denoted by @xmath89 , and the full collection of sides of @xmath0-cells of @xmath6 which are present at @xmath47-cells of @xmath6 will be denoted by @xmath90 . notice that saying a side @xmath88 is present at @xmath80 is equivalent to saying that the map @xmath51 extends to a map @xmath91 . notice also that if @xmath82 and @xmath92 are distinct @xmath47-cells of @xmath50 which are mapped to the same @xmath47-cell @xmath80 of @xmath6 , then @xmath93 and @xmath94 are distinct sides at @xmath80 , even though @xmath82 and @xmath92 come from the same @xmath0-cell @xmath50 and are mapped to the same @xmath47-cell @xmath80 . thus a @xmath0-cell @xmath50 whose boundary has length @xmath26 will have exactly @xmath26 distinct sides in @xmath90 . alternatively , the elements of @xmath89 can be viewed as the connected components in @xmath95 , where @xmath96 is a small open ball around a point in the interior of @xmath80 . next , let @xmath46 be a map , let @xmath88 be a side of @xmath6 which is present at @xmath80 , and let @xmath81 be a @xmath47-cell of @xmath4 with @xmath97 . we say that the side @xmath98 is _ present at @xmath81 _ if the map @xmath98 factors through a map @xmath99 as indicated in the following commutative diagram : @xmath100 specifically , there must exist a map @xmath101 such that @xmath102 is the map @xmath98 . if the map @xmath93 does not factor through @xmath58 then @xmath88 is said to be _ missing at @xmath81_. the set of all sides of @xmath6 which are present at @xmath81 will be denoted by @xmath103 , while the set of all sides of @xmath6 which are missing at @xmath81 will be denoted @xmath104 . [ rem : sides ] it is important to notice that the definitions of the sets @xmath103 and @xmath89 both refer to the sides of @xmath0-cells of the complex @xmath6 . in particular , if @xmath105 then @xmath106 , and @xmath107 is not comparable with either of these since it is a subset of @xmath108 . moreover , @xmath103 can be smaller than @xmath107 if the map @xmath58 is not a near - immersion . in fact , @xmath58 is a near - immersion if and only if @xmath109 for all @xmath47-cells @xmath110 . [ def : perimeter ] let @xmath46 be a combinatorial map between @xmath0-complexes . we define the _ unit perimeter _ of @xmath58 to be @xmath111 for each @xmath47-cell @xmath81 of @xmath4 , we can either count the sides of @xmath6 at @xmath112 which are missing at @xmath81 , or else we can count the number of sides of @xmath6 that are present at @xmath80 and then subtract off those which are also present at @xmath81 . from the first description it is clear that the perimeter of @xmath58 is nonnegative . the following examples will illustrate these distinctions . in particular , they will illustrate the significance of the maps @xmath46 and @xmath113 , respectively . [ exmp : cells ] let @xmath6 be the complex formed by attaching two squares along a common @xmath47-cell @xmath80 , and let @xmath4 be another complex which is isomorphic to @xmath6 with common @xmath47-cell @xmath81 . let @xmath114 be an isomorphism , and let @xmath115 be a map which sends @xmath81 to @xmath80 but which folds the two squares of @xmath4 to the same square of @xmath6 . observe that @xmath116 but @xmath117 . [ exmp : sides ] let @xmath6 be the standard @xmath0-complex of the presentation @xmath118 and let @xmath51 be the unique @xmath0-cell of @xmath6 . notice that @xmath87 wraps three times around the path @xmath119 in @xmath6 , and that there are exactly six sides present at the @xmath47-cell labeled @xmath120 and exactly three sides present at the @xmath47-cell labeled @xmath121 . inside the universal cover @xmath122 of @xmath6 , one can find three distinct @xmath0-cells which share the same boundary cycle . let @xmath123 be the union of two of these three @xmath0-cells and define @xmath124 to be the composition @xmath125 . observe that @xmath4 is a sphere and @xmath58 is an immersion . if @xmath81 is a @xmath47-cell labeled @xmath121 in @xmath4 and @xmath80 is its image under @xmath58 , then @xmath126 , @xmath127 , and @xmath128 . thus @xmath129 . to see the importance of the map @xmath51 , let @xmath82 be a @xmath47-cell labeled @xmath121 in @xmath50 and let @xmath98 be the corresponding map of pairs . there is a @xmath47-cell @xmath81 in @xmath4 such that @xmath88 is missing at @xmath81 even though there are two distinct maps from @xmath88 to @xmath4 which send @xmath82 to @xmath81 and which when composed with @xmath58 agree with the map @xmath51 on @xmath87 . the side @xmath88 is missing from @xmath81 because neither of these maps agree with @xmath51 on the interior of @xmath50 . [ exmp : zzz ] let @xmath130 = [ a , c ] = [ b , c ] = 1 \rangle$ ] be the standard presentation of the free abelian group on three generators and let @xmath6 be the standard @xmath0-complex corresponding to this presentation . the universal cover @xmath122 of @xmath6 is usually thought of as the points of @xmath131 with @xmath132 or @xmath133 or @xmath134 . that is , @xmath122 is isomorphic to the union of the integer translates of the @xmath80-@xmath81 , @xmath81-@xmath135 , and @xmath80-@xmath135 planes . the @xmath47-cells and @xmath0-cells of @xmath122 are unit intervals and unit squares . moreover , @xmath122 is labeled so that @xmath47-cells parallel to the @xmath80-axis are labeled by the generator @xmath120 and directed in the positive @xmath80 direction . similarly , @xmath47-cells parallel to the @xmath81-axis and @xmath135-axis are labeled by @xmath121 and @xmath136 directed in the positive @xmath81-direction and @xmath135-direction respectively . of example [ exmp : zzz ] [ fig : zzz ] ] if @xmath4 is a 1-by-1-by-1 box with four walls , a bottom and no top ( see figure [ fig : zzz ] ) , and @xmath58 is the obvious embedding of @xmath4 into @xmath122 , then the perimeter of @xmath4 is as follows . each @xmath47-cell @xmath55 along the top of the open box contributes a perimeter of @xmath1 corresponding to the sides of the three squares which are incident at @xmath137 in @xmath122 but which do not lift to @xmath4 . the vertical @xmath47-cells along the sides of the box contribute a perimeter of @xmath0 each , as do the @xmath47-cells on the bottom of the box . the total perimeter of @xmath4 is @xmath138 sides of squares . next , consider a more typical example of a map which is not an embedding . let @xmath3 denote the composition map @xmath139 . the exact same count shows that the perimeter of @xmath140 is also @xmath138 . we will now define a more flexible notion of perimeter which employs a weighting on the sides of the @xmath0-cells in @xmath6 . the weighted perimeter of @xmath46 is intuitively just the sum of the weights of the corresponding missing sides . we will now make this notion more precise . [ def : weighted - perimeter ] a _ weight function _ on a @xmath0-complex @xmath6 is a function of the form @xmath141 . for most of our applications we will require that the weight of a side be nonnegative . let @xmath141 be a weight function on @xmath6 , and let @xmath46 be a combinatorial map of @xmath0-complexes . the _ weighted perimeter _ of @xmath46 is the sum of the weights of the sides of @xmath6 which are missing at @xmath47-cells of @xmath4 . more precisely , the weighted perimeter is defined to be the following double sum : @xmath142 notice that the weighted perimeter is equivalent to the unit perimeter when each side is assigned a weight of @xmath47 . this weight function will be called the _ unit weighting_. note also that if the assigned weights are nonnegative , then the weighted perimeter of @xmath58 will be nonnegative . when the map @xmath58 is understood we will write @xmath5 or @xmath143 for @xmath144 . a @xmath0-complex @xmath6 is a _ weighted @xmath0-complex _ if each of the sides of @xmath6 has been assigned a nonnegative integer weight , the perimeter of each @xmath47-cell is finite , and the weight of each @xmath0-cell is positive . although the definition of a weighted @xmath0-complex adopts the requirement of an integer weighting , it is sufficient , and often quite natural , to use a finite set of rationals . in the theorems which follow , a successful real - valued weight function can always be approximated by a successful rational weight function . after clearing the denominators , we then would obtain a successful integer - valued weight function . thus there is no real loss of generality in assuming that the values of the weights are integers . perimeters of weighted @xmath0-complexes satisfy the following useful property : [ lem : surj ] let @xmath6 be a weighted @xmath0-complex and consider maps @xmath145 , @xmath46 , and @xmath146 . if @xmath147 is surjective then @xmath148 . first notice that @xmath149 whenever @xmath150 ( or equivalently that @xmath151 ) . since @xmath147 is surjective , for every @xmath47-cell @xmath81 in @xmath4 we can select a @xmath47-cell @xmath135 in @xmath66 with @xmath150 . it is also clear that the @xmath135 s chosen for distinct @xmath81 s are themselves distinct . thus the terms in the sum for @xmath152 can be identified with distinct terms in the sum for @xmath153 . finally , since the weights are nonnegative it follows that @xmath154 . given a weight function on @xmath6 there is also an induced value assigned to each of the @xmath47-cells and @xmath0-cells of @xmath6 . we define the _ perimeter of a @xmath47-cell _ @xmath80 in @xmath6 to be the sum of the weights assigned to the sides in @xmath89 . this agrees with our earlier definition of perimeter in the sense that it is the weighted perimeter of the map @xmath155 which sends the single @xmath47-cell to @xmath80 in @xmath6 . in particular , it measures the weights of the sides which are _ not _ present when the @xmath47-cell @xmath80 is considered in isolation . we define the _ weight of a @xmath0-cell _ @xmath50 in @xmath6 to be the sum of the weights assigned to the sides of the form @xmath88 for some @xmath82 in @xmath87 . the sum of the weights of the sides of a @xmath0-cell , on the other hand , is called a weight since it is the sum of the weights of sides which are present in the @xmath0-cell itself and it ignores the weights of the other sides which are incident at @xmath47-cells in its boundary . formally , we have the equations @xmath156 @xmath157 if @xmath4 is compact and the map @xmath114 is a near - immersion , then the perimeter of @xmath4 can be calculated from the perimeters of its @xmath47-cells and the weights of its @xmath0-cells . specifically we have the following result . [ lem : perimeter - rewritten ] if @xmath6 is a weighted @xmath0-complex , @xmath4 is compact , and the map @xmath114 is a near - immersion , then @xmath158 the second summation is the one which requires the immersion hypothesis . by remark [ rem : sides ] , the restriction on @xmath58 implies that the sides of @xmath6 which are present at @xmath81 are in one - to - one correspondence with the sides of @xmath4 which are present at @xmath81 . these sides of @xmath4 can then be collected together according to the @xmath0-cell in @xmath4 to which they belong , and then the sum of the weights of the sides of a particular @xmath0-cell @xmath159 in @xmath4 can be rewritten as the weight of the @xmath0-cell in @xmath6 which is the image of @xmath159 under @xmath58 . the following example illustrates these types of calculations . [ exmp : weighted zzz ] let @xmath6 , @xmath4 , and @xmath160 be the spaces and maps described in example [ exmp : zzz ] . specifically , let @xmath6 denote the standard @xmath0-complex of the presentation @xmath161 denote the three @xmath0-cells of @xmath6 by @xmath162 , @xmath163 and @xmath164 , and observe that each of them has four sides . since the four letters of each defining word are in one - to - one correspondence with the four sides of the @xmath0-cell , the weights of the sides of the @xmath0-cell can be indicated by a sequence of four numbers . if we assign weights to the sides of @xmath0-cells of @xmath6 via the sequences @xmath165 , and @xmath166 , then the reader can verify that @xmath167 , @xmath168 , @xmath169 , @xmath170 , @xmath171 , and @xmath172 . since the map @xmath173 is an immersion , by lemma [ lem : perimeter - rewritten ] the weighted perimeter of the map @xmath114 can be calculated as follows : @xmath174 in this section we describe a general framework for showing that groups are coherent and then employ the notion of perimeter to state our main hypothesis and to prove our main coherence theorem . [ def : complexity - fns ] let @xmath6 be a fixed @xmath0-complex , let @xmath175 be a well - ordered set , and let @xmath176 be a function which assigns an element of @xmath33 to each map @xmath46 with a compact domain . the function @xmath176 is a _ complexity function _ for @xmath6 , and the value @xmath177 is the _ complexity of the map @xmath58_. in practice , @xmath175 will either be @xmath178 with the usual ordering , or @xmath33 will be @xmath179 and @xmath180 is defined so that @xmath181 if either @xmath182 or @xmath183 and @xmath184 . this is the usual lexicographic ordering on ordered pairs . [ reduction - method ] let @xmath6 be a fixed @xmath0-complex and let @xmath176 be a complexity function for @xmath6 . if for all compact spaces @xmath4 and maps @xmath160 such that @xmath58 is not already @xmath7-injective , there is a `` procedure '' ( in any sense of the word ) which produces a compact space @xmath66 and a map @xmath185 such that @xmath186 and such that @xmath187 , then this procedure will be called a _ reduction method _ for @xmath176 . [ rem : compact ] notice that for every finitely generated subgroup @xmath18 in @xmath14 there exists a compact space @xmath4 and map @xmath114 such that the image of @xmath188 under @xmath62 is exactly @xmath18 . one procedure for creating @xmath4 and @xmath58 goes as follows : suppose that @xmath18 is generated by @xmath26 elements of @xmath14 and represent each of these generators by a closed path in the @xmath47-skeleton of @xmath6 starting at the basepoint . next let @xmath4 be a bouquet of @xmath26 circles , and after subdividing @xmath4 , define @xmath114 so that the restriction of @xmath58 to the @xmath40-th subdivided circle is identical to the @xmath40-th closed path . alternatively , we could let @xmath8 be the based covering space of @xmath6 corresponding to the subgroup @xmath18 , and let @xmath4 be the union of the based lifts to @xmath8 of a finite set of closed based paths representing the generators of @xmath18 in @xmath14 . it is clear that both of these constructions yield @xmath7-surjective maps . the latter has the advantage of being an immersion . the following theorem is the philosophical basis for the coherence results in this paper . [ thm : complexity=>coherent ] let @xmath6 be a fixed space and let @xmath176 be a complexity function for @xmath6 . if there is a reduction method for @xmath176 then @xmath14 is coherent . let @xmath18 be an arbitrary finitely generated subgroup of @xmath14 . by remark [ rem : compact ] , there is at least one combinatorial map @xmath46 such that @xmath4 is compact and @xmath189 . if @xmath58 is not @xmath7-injective then there is another combinatorial map with the same properties which has a strictly lower complexity . since @xmath175 is well ordered , there can not be an infinite sequences of reductions . hence the process of replacing one combinatorial map with another must terminate at a @xmath7-injective combinatorial map @xmath190 where @xmath66 is compact and @xmath191 . since @xmath147 is @xmath7-injective , @xmath192 is itself isomorphic to @xmath18 . since @xmath66 is compact , @xmath18 is finitely presented . in particular , a standard @xmath0-complex for a finite presentation of @xmath18 can be obtained by contracting a maximal tree in @xmath193 . note that if the reduction method for @xmath176 is constructive , then the proof of theorem [ thm : complexity=>coherent ] can be used as an algorithm to effectively compute finite presentations for finitely generated subgroups . many of the reduction methods we introduce are in fact constructive and in section [ sec : algorithms ] we explicitly describe a resulting algorithm . [ coherence=>complexity ] the converse of theorem [ thm : complexity=>coherent ] is also true in the following sense . given a space @xmath6 with a coherent fundamental group , we define the complexity of a map @xmath46 where @xmath4 is compact to be the minimum number of @xmath0-cells which must be added to @xmath4 to yield a @xmath7-injection . it is easy to see that this is indeed a complexity function , that the `` procedure '' of adding one of the necessary @xmath0-cells is a method of reducing the complexity , and that there can be no infinite sequences of reductions . we will now specialize to the case where weighted perimeter is used to measure the complexity of a map . [ def : gprh ] let @xmath6 be a weighted @xmath0-complex . it will satisfy the _ perimeter reduction hypothesis _ if for any compact and connected space @xmath4 and for any based @xmath47-immersion @xmath46 which is not @xmath7-injective , there exists a based map @xmath194 and a commutative diagram @xmath195 such that @xmath196 is compact and connected , @xmath197 and @xmath198 has the same @xmath7-image as @xmath3 . typically , the final requirement that @xmath199 is deduced from a more stringent requirement that @xmath200 is @xmath7-surjective . [ thm : gct ] let @xmath6 be a weighted @xmath0-complex which satisfies the perimeter reduction hypothesis . \a ) if @xmath4 is a compact connected subcomplex of a cover @xmath8 of @xmath6 , and the inclusion @xmath9 is not @xmath7-injective , then @xmath4 is contained in a compact connected subcomplex @xmath10 such that @xmath11 . \b ) for any compact subcomplex @xmath12 , there exists a compact connected subcomplex @xmath4 containing @xmath13 , such that @xmath5 is minimal among all such compact connected subcomplexes containing @xmath13 . consequently @xmath14 is coherent . to prove statement a , suppose that @xmath4 is connected and compact but the inclusion map @xmath9 is not @xmath7-injective . then by the perimeter reduction hypothesis , there exists a commutative diagram @xmath195 such that @xmath196 is compact and connected , such that @xmath197 , and such that @xmath198 has the same @xmath7-image as @xmath3 . observe that @xmath198 lifts to a map @xmath201 which extends the lift of @xmath4 to @xmath8 . let @xmath10 denote the image of @xmath196 in @xmath202 . by lemma [ lem : surj ] , @xmath203 . statement b follows immediately from the fact that the perimeters of compact subcomplexes containing @xmath13 are nonnegative integers . to see that @xmath14 is coherent , let @xmath8 be a based cover of @xmath6 such that @xmath16 is finitely generated . observe that there exists a based compact connected subspace @xmath12 whose inclusion induces a @xmath7-surjection . let @xmath204 denote a compact connected subspace containing @xmath13 such that @xmath5 is minimal among all such compact connected subspaces . then @xmath9 is @xmath7-surjective because @xmath205 , and @xmath9 is @xmath7-injective by statement a. in the remainder of the article we will provide three conditions which will imply the perimeter reduction hypothesis : the @xmath0-cell reduction hypothesis , the path reduction hypothesis , and the fan reduction hypothesis . these three hypotheses are more concrete than the perimeter reduction hypothesis and thus tend to be more useful in establishing the coherence of specific presentations . we conclude this section by noting that theorem [ thm : gct ] very nearly shows that finitely generated covers of complexes satisfying the perimeter reduction hypothesis have compact cores . [ def : core ] a subcomplex @xmath4 of the complex @xmath66 is a _ core _ of @xmath66 if the inclusion map @xmath206 is a homotopy equivalence . since @xmath4 and @xmath66 are cw - complexes , @xmath4 is a core of @xmath66 if and only if there is a strong deformation retraction from @xmath66 to @xmath4 , which is true if and only if the map @xmath207 induces an isomorphism on all of the homotopy groups ( @xcite ) . note that when @xmath66 is an aspherical @xmath0-complex , @xmath4 will be a core for @xmath66 if and only if the inclusion of @xmath4 induces a @xmath7-isomorphism . indeed , if @xmath208 is @xmath7-injective , then the based component of the preimage of @xmath4 in the universal cover @xmath209 is clearly isomorphic to the universal cover @xmath210 of @xmath4 . but then @xmath211 , and so we see that @xmath210 and thus @xmath4 is aspherical . [ thm : compactcore ] let @xmath6 be a weighted aspherical @xmath0-complex which satisfies the perimeter reduction hypothesis . if @xmath15 is a covering space and @xmath16 is finitely generated , then every compact subcomplex of @xmath8 is contained in a compact core of @xmath8 . this follows immediately from theorem [ thm : gct ] and definition [ def : core ] . the existence of a compact core in a @xmath0-complex is a nontrivial fact . for example , there exists a covering space @xmath8 of a @xmath0-complex @xmath6 with a single @xmath0-cell , such that @xmath16 is finitely generated , but @xmath8 has no compact core . see @xcite for details . the restriction in theorem [ thm : compactcore ] that @xmath6 be aspherical is not particularly stringent since one of our main sources of applications will be small cancellation complexes , and small cancellation complexes in which none of the @xmath0-cells are attached by proper powers are known to be aspherical @xcite . roughly speaking , a @xmath0-cell @xmath50 is attached by a proper power if @xmath54 is obtained by traversing a closed path in @xmath6 two or more . see definition [ def : exponent ] . it appears likely that if @xmath6 is a compact @xmath0-complex which satisfies the perimeter reduction hypothesis then @xmath14 acts properly discontinuously on a contractible @xmath0-complex . we have been unable to decide whether this is the case . by theorem [ thm : gct ] , the fundamental group of a weighted @xmath0-complex @xmath6 is coherent if there is a method for reducing the perimeter of the maps @xmath140 which are not @xmath7-injections . one of the simplest possibilities is where the perimeter is reduced through the attachment of a single @xmath0-cell , a possibility which will be examined in detail in section [ sec:2cell ] . in this section , we provide the definitions and results about paths , @xmath0-cells and attachments which will be needed . we will now describe the two elementary ways of changing a path @xmath212 : to remove a backtrack and to push across a @xmath0-cell . [ def : removing - backtracks ] if @xmath213 contains a subpath of the form @xmath214 where @xmath55 is a @xmath47-cell of @xmath6 , then there is another path @xmath215 obtained by simply removing these two @xmath47-cells from the path . such a change is called _ removing a backtrack_. notice that the paths @xmath68 and @xmath215 are homotopic relative to their endpoints , that a path @xmath69 is immersed if and only if it has no backtracks to remove , and that removing a backtrack reduces the length of the path . let @xmath51 be a @xmath0-cell , and let @xmath73 be a subpath of @xmath87 . there exists a unique subpath @xmath159 of @xmath216 , called the _ complement of @xmath73 in @xmath50 _ , such that the concatenation @xmath217 is a closed path which corresponds to the boundary cycle @xmath87 . note that if @xmath218 , then @xmath159 is a trivial path . [ def : pushing - across ] let @xmath213 be a path , let @xmath113 be a @xmath0-cell , and let @xmath73 be a subpath of both @xmath69 and @xmath87 , so that we have the following commutative diagram : @xmath219 let @xmath159 be the complement of @xmath73 in @xmath50 , and observe that since @xmath159 and @xmath73 have the same endpoints in @xmath6 , we can form a new path @xmath220 by substituting @xmath159 for the subpath @xmath73 of @xmath69 . in particular , if the path @xmath68 is the concatenation of a path @xmath221 followed by the path @xmath73 followed by a path @xmath222 , then the modified path @xmath215 is the concatenation @xmath223 . the replacement of @xmath224 by @xmath225 is called _ pushing across the @xmath0-cell @xmath113_. it is clear that if @xmath220 is obtained from @xmath69 by pushing across a @xmath0-cell , then @xmath69 and @xmath220 will be homotopic relative to their endpoints . notice also that @xmath226 whenever @xmath227 . [ def : exponent ] let @xmath6 be a @xmath0-complex , and let @xmath113 be one of its @xmath0-cells . let @xmath26 be the largest number such that the map @xmath228 can be expressed as a path @xmath229 in @xmath6 , where @xmath23 is a closed path in @xmath6 . this number @xmath26 , which measures the periodicity of the map of @xmath230 , is the _ exponent _ of @xmath50 , and a path such as @xmath23 is a _ period _ for @xmath87 . notice that any other closed path which determines the same cycle as @xmath23 will also be a period of @xmath87 . if the exponent @xmath26 is greater than @xmath47 , then the @xmath230 is called a _ proper power_. [ def : packets ] let @xmath50 be a @xmath0-cell in @xmath6 of exponent @xmath26 and let @xmath23 be a period of @xmath87 . the attaching map @xmath230 can be expressed as a path @xmath231 . consider a circle subdivided into @xmath232 @xmath47-cells , and attach a copy of @xmath50 by wrapping @xmath216 around the circle @xmath26 times . we call the resulting @xmath0-complex @xmath233 . note that there is a map @xmath234 such that @xmath51 factors as @xmath235 . observe that @xmath236 and that the universal cover of @xmath233 has a @xmath47-skeleton which is identical to that of @xmath50 together with @xmath26 distinct copies of @xmath50 attached by embeddings . the universal cover of @xmath233 is the _ packet of @xmath50 _ and is denoted by @xmath237 . technically we should write @xmath238 but we will use the notation of @xmath239 since @xmath50 is its own universal cover and thus there is no danger of confusion . notice that if the exponent of @xmath50 is @xmath47 then the packet @xmath239 is the same as @xmath50 itself . notice also that the map @xmath240 can be viewed as an extension of the map @xmath113 . let @xmath46 be a fixed map . the map @xmath58 will be called _ packed _ if whenever there is a lift of a @xmath0-cell @xmath241 to a @xmath0-cell @xmath242 , there is also a lift of @xmath240 to a map @xmath243 which extends the map @xmath242 . since we will treat the packets @xmath239 as the basic building blocks of our @xmath0-complexes , almost all of the maps under discussion will be packed . [ def : attachments ] let @xmath244 be an arbitrary packed map and let @xmath241 be a @xmath0-cell in @xmath6 . the pair of paths @xmath245 will be called a _ @xmath0-cell attachment site _ if they satisfy the following conditions : 1 . the path @xmath246 is a subpath of @xmath87 2 . the diagram @xmath247commutes 3 . there does not exist a map @xmath248 which is a lift of the map @xmath249 such that the composition @xmath250 equals the path @xmath251 . intuitively , a @xmath0-cell attachment site is a portion of the boundary of @xmath50 which is found in the complex @xmath4 at a location where the packet @xmath239 does not already exist . in other words , it is a place at which attaching a copy of @xmath239 will have an effect on the perimeter of the map . notice that when the length of @xmath73 is equal to the length of @xmath87 , the path @xmath251 may have distinct endpoints even though the endpoints of the path @xmath252 are identical . [ def : maximal - attachments ] a @xmath0-cell attachment site is _ maximal _ if there does not exist another pair of maps @xmath253 where @xmath254 is a proper subpath of @xmath255 and @xmath256 is a proper subpath of @xmath257 . technically , we require that there does not exist a proper inclusion @xmath258 such that @xmath259 is the map @xmath246 and @xmath260 is the map @xmath251 . this forces @xmath73 to appear as a proper subpath of @xmath261 in the same manner in both cases . if @xmath262 and the @xmath0-cell attachment site is maximal we will call it an _ incomplete @xmath0-cell attachment_. when @xmath263 , we will call this a _ complete @xmath0-cell attachment_. notice that complete attachments are automatically maximal . let @xmath6 be a weighted @xmath0-complex and let @xmath140 be a packed map . a @xmath0-cell attachment @xmath264 will be called a _ @xmath0-cell perimeter reduction _ if @xmath265 . if @xmath266 it will be called a _ weak @xmath0-cell perimeter reduction_. [ rem : peri - reduction ] a @xmath0-cell perimeter reduction @xmath264 is so named because it can be used to reduce the perimeter of the map @xmath46 ( lemma [ lem : attachments ] ) . the main idea is as follows : simply attach the packet @xmath239 to @xmath4 along the path @xmath73 . technically , the identification space @xmath267 is formed by identifying the image of the @xmath47-cells of @xmath73 in @xmath239 with their image in @xmath4 . for simplicity , we write @xmath268 for the resulting complex , and we call the extended map @xmath269 . since @xmath265 , the cells which are in @xmath239 and not in @xmath73 must make a net negative contribution to the perimeter and consequently @xmath270 . the details and the qualifications which are necessary to justify this calculation are contained in lemma [ lem : complete ] and lemma [ lem : incomplete ] . these two technical lemmas will be the key ingredients in the proof of theorem [ thm:2cellct ] . the following lemma shows how the relationship between the perimeter of @xmath73 and the perimeter of the packet @xmath239 can be reformulated as a relationship between the exponent of @xmath50 , the weight of @xmath50 , and the perimeter of the complement of @xmath73 . this alternative form makes it easier to verify that a specific reduction is a perimeter reduction . the original form is easier to understand conceptually . [ lem : s - form ] let @xmath6 be a weighted @xmath0-complex , let @xmath113 be a @xmath0-cell , and let @xmath246 be a subpath of @xmath87 . if @xmath26 is the exponent of @xmath50 and @xmath159 is the complement of @xmath73 in @xmath50 , then @xmath271 . consequently , @xmath272 if and only if @xmath273 , and the first inequality is strict if and only if the second one is strict . since the map @xmath240 is an immersion , lemma [ lem : perimeter - rewritten ] can be used to yield the first equation . the inequalities then follow as simple rearrangements of this basic equation . we conclude this section with the notion of a redundant @xmath0-cell . let @xmath46 be a fixed map and let @xmath274 and @xmath275 be @xmath0-cells in @xmath4 . we say that @xmath276 is _ redundant ( relative to @xmath162 and the map @xmath3 ) _ provided that @xmath162 and @xmath163 are distinct @xmath0-cells in @xmath4 which have the same boundary cycle , but @xmath162 and @xmath163 project to the same @xmath0-cell in @xmath6 . more precisely , their interiors in @xmath4 are disjoint , but there exists a map @xmath277 which restricts to @xmath278 , such that the following two diagrams commute : @xmath279 because of the way that perimeter is calculated , redundant @xmath0-cells have no effect on the perimeter of @xmath140 . this is made precise below and will be used in the proofs in section [ sec:2cell ] . [ lem : redundant ] let @xmath6 be a weighted @xmath0-complex , let @xmath140 be a map , and let @xmath162 and @xmath163 be redundant @xmath0-cells of @xmath4 . if @xmath10 is @xmath4 minus the interior of @xmath162 and @xmath280 is the restriction of @xmath58 to @xmath10 , then @xmath281 . more generally , if @xmath4 and @xmath10 differ by the addition or removal of redundant @xmath0-cells , then @xmath282 . since the @xmath47-skeletons are identical and @xmath283 , it is clear that each side that is missing at @xmath81 in @xmath4 is also missing at @xmath81 in @xmath10 . to see the reverse implication , let @xmath98 be a side of @xmath6 which is present at @xmath81 in @xmath4 . if @xmath91 lifts to a side of @xmath81 which is a side of the @xmath0-cell @xmath284 , then by the definition of redundant @xmath0-cells , it also lifts to a side of the @xmath0-cell @xmath275 at @xmath81 . thus every side at @xmath80 which is present at @xmath81 in @xmath4 is also present at @xmath81 in @xmath10 . the final assertion is now immediate . finally , we relate the lack of redundant @xmath0-cells to immersions in the following lemma whose proof is immediate . [ lem : no - redundant->immersion ] if @xmath46 is a @xmath47-immersion and @xmath4 has no redundant @xmath0-cells , then @xmath58 is an immersion . in this section we show how @xmath0-cell perimeter reductions can be used to lower the perimeter of a map @xmath140 . at the end of the section we use this to prove a @xmath0-cell version of our coherence theorem . [ lem : complete ] let @xmath6 be a weighted @xmath0-complex , let @xmath46 be a packed @xmath47-immersion , and suppose that @xmath5 is finite . if @xmath264 is a complete @xmath0-cell attachment , then the perimeter of the induced map @xmath285 satisfies the equation @xmath286 since by assumption @xmath263 , the space @xmath287 can be formed by first identifying the endpoints of @xmath73 in @xmath4 , if they are not already identical , and then attaching the packet @xmath239 along its boundary . next , since the space @xmath4 , with the two endpoints of @xmath73 identified , is a subcomplex of @xmath196 with an identical @xmath47-skeleton , any side of @xmath6 which is missing at @xmath81 in @xmath196 is also missing at @xmath81 in @xmath4 . this shows that the terms in the sum defining @xmath288 are contained as distinct terms in the sum defining @xmath5 . let @xmath88 be a side of @xmath6 , and let @xmath81 be the image of this @xmath47-cell @xmath82 under the map @xmath251 . if the side @xmath88 was already present at @xmath81 , then , using the fact that @xmath46 is a packed @xmath47-immersion , we find that there already existed a lift of @xmath240 to @xmath4 for which the composition @xmath289 is the given map @xmath251 . since this contradicts our assumption that @xmath264 is a @xmath0-cell attachment site , we have shown that the side @xmath88 was missing at @xmath81 in @xmath4 , even though it is clearly present at @xmath81 in @xmath196 . if we repeat this argument for each of the sides of @xmath50 we can conclude that @xmath290 , which is less than @xmath5 since @xmath291 . a careful argument would show that @xmath292 where @xmath293 is the exponent of the @xmath0-cell @xmath294 . [ lem : incomplete ] let @xmath6 be a weighted @xmath0-complex , let @xmath46 be a packed @xmath47-immersion and suppose that @xmath5 is finite . if @xmath264 is an incomplete @xmath0-cell attachment then the perimeter of the induced map @xmath295 satisfies the equation @xmath296 since the perimeter of @xmath46 is unaffected by the addition or removal of redundant @xmath0-cells from @xmath4 ( lemma [ lem : redundant ] ) , we might as well assume that @xmath4 has no redundancies . by lemma [ lem : no - redundant->immersion ] this means that we may assume that @xmath58 is an immersion . we will now show that the map @xmath297 is a near - immersion . since the maps @xmath140 and @xmath240 are immersions we only need to show that this is true when @xmath81 lies in the image of @xmath73 under the map @xmath298 . let @xmath88 be a side of @xmath6 which is present at @xmath81 in @xmath196 . if this side was already present at @xmath81 in @xmath4 , then , using the fact that @xmath299 is a packed immersion , we find that there already existed a lift of @xmath240 to @xmath4 for which the composition @xmath300 is the map @xmath256 . since this contradicts our assumption that @xmath301 is a @xmath0-cell attachment , we conclude that @xmath88 must be missing at the @xmath47-cell @xmath81 in @xmath4 . next , suppose that @xmath88 is a side of @xmath6 which is present in @xmath239 at @xmath302 and present in @xmath239 at @xmath303 . suppose further that both @xmath302 and @xmath303 lie in @xmath73 and that they are sent to the same @xmath47-cell @xmath81 in @xmath196 . since all of the sides of @xmath50 are distinct , the only way in which this could happen is if the exponent of @xmath50 is nontrivial , these two copies of @xmath50 in @xmath239 are distinct , and the @xmath47-cells @xmath302 and @xmath303 differ by a path which is a multiple of the period @xmath23 of @xmath87 . as a consequence we find that the path from @xmath302 to @xmath303 in @xmath239 is sent to a closed path in @xmath4 which is a multiple of a period of @xmath87 , and it is possible to extend the path @xmath251 to the entire boundary of @xmath50 , thereby contradicting the maximality assumption on @xmath73 . we thus conclude that distinct sides of @xmath0-cells in @xmath239 are sent to distinct sides of @xmath0-cells in @xmath196 . since we also showed that these sides are disjoint from the sides of @xmath6 which are present at @xmath81 in @xmath4 , we now know that the map from @xmath196 to @xmath6 is an immersion in a small neighborhood of a point in the interior of each @xmath47-cell . if we assume for the moment that @xmath4 is compact , then we can calculate the perimeter of @xmath196 using equation ( [ eq : weight - stuff ] ) of lemma [ lem : perimeter - rewritten ] . according to equation ( [ eq : weight - stuff ] ) , the perimeter of @xmath196 equals the weight of its @xmath47-cells minus the weight of its @xmath0-cells . if we apply equation ( [ eq : weight - stuff ] ) to @xmath239 and @xmath4 separately then we add the weight of their @xmath47-skeletons and subtract the weights of their @xmath0-cells . the difference between these counts is precisely the @xmath47-cells of @xmath73 in @xmath239 which get identified to @xmath47-cells in @xmath4 in the space @xmath196 . this proves equation ( [ eq : incomplete ] ) . in the general case where we assume that @xmath5 is finite but not that @xmath4 is compact , then we can not use equation ( [ eq : weight - stuff ] ) as we did above . instead we argue as follows : let @xmath159 be the complement of @xmath73 in @xmath87 . the change in perimeter from @xmath4 to @xmath196 can be computed by first adding @xmath304 corresponding to the new @xmath47-cells in @xmath196 and then subtracting @xmath305 corresponding to the new sides . the resulting change in perimeter is @xmath306 , which is equal to @xmath307 by lemma [ lem : s - form ] . combining lemma [ lem : complete ] and lemma [ lem : incomplete ] , we have the following . [ lem : attachments ] let @xmath6 be a weighted @xmath0-complex , let @xmath46 be a packed @xmath47-immersion and suppose that @xmath5 is finite . if @xmath264 is a @xmath0-cell perimeter reduction then the perimeter of the induced map @xmath285 is strictly less than @xmath5 . if it is a weak @xmath0-cell perimeter reduction , @xmath308 . without loss of generality we may assume that the @xmath0-cell perimeter reduction @xmath264 is maximal . let @xmath309 . that @xmath270 [ @xmath308 ] now follows immediately from either lemma [ lem : complete ] or lemma [ lem : incomplete ] depending on whether the reduction is complete or incomplete . in addition to the process of attaching @xmath0-cells , we will also need a second operation called folding . [ def : folding ] let @xmath3 be a map between @xmath0-complexes and let @xmath311 be a length @xmath0 path whose projection to @xmath6 is of the form @xmath214 ( i.e. a backtrack ) . if the @xmath47-cells of @xmath69 are distinct in @xmath4 , then the map @xmath140 can be factored as @xmath312 where the complex @xmath10 is obtained from @xmath4 by identifying the endpoints of @xmath69 ( if they are not already identical ) and then identifying the @xmath47-cells in the image of @xmath311 in the obvious way . the complex @xmath10 is said to be obtained from @xmath4 by _ folding along the path @xmath69_. if @xmath4 can be folded along some path @xmath311 , then @xmath3 _ admits a fold_. [ def : prh ] a space @xmath6 is said to satisfy the _ @xmath0-cell reduction hypothesis _ if for any map @xmath46 which is a packed @xmath47-immersion which is not a @xmath7-injection , there exists a @xmath0-cell @xmath51 and a @xmath0-cell perimeter reduction @xmath301 . notice that if @xmath6 satisfies the @xmath0-cell reduction hypothesis and @xmath140 is a packed map which does not admit a fold or a @xmath0-cell perimeter reduction , then the induced map @xmath313 is injective . [ thm:2cellct ] if @xmath6 is a weighted @xmath0-complex that satisfies the @xmath0-cell reduction hypothesis , then it satisfies the perimeter reduction hypothesis , and thus @xmath14 is coherent . let @xmath140 be a @xmath47-immersion which is not @xmath7-injective . since adding the @xmath0-cells necessary to make @xmath3 a packed map does not increase perimeter , we may assume it is packed without loss of generality . by hypothesis , there is a @xmath0-cell perimeter reduction and by lemma [ lem : attachments ] the perimeter of @xmath196 will be smaller . the fact that @xmath4 and @xmath196 have the same @xmath7 image in @xmath6 is obvious . that @xmath14 is coherent now follows from theorem [ thm : gct ] . the @xmath0-cell coherence theorem ( theorem [ thm:2cellct ] ) can also be presented as an algorithm for constructing finite presentations from a given finite set of generators . the algorithm may be viewed as a generalization of stallings algorithm for graphs @xcite . [ thm : algorithm ] if @xmath6 is a compact weighted @xmath0-complex which satisfies the @xmath0-cell reduction hypothesis , then there is an algorithm which produces a finite presentation for any subgroup of @xmath14 given by a finite set of generators . to help clarify that the algorithm terminates , we will use a complexity function other than the usual perimeter . we define the complexity of a map @xmath114 to be the ordered pair @xmath314 where @xmath5 is the perimeter of the map , @xmath315 is the number of @xmath47-cells in @xmath4 , and the ordering is the dictionary ordering . let @xmath18 be a subgroup of @xmath14 generated by a set of @xmath82 elements represented by closed based paths . we let @xmath316 be a based bouquet of @xmath82 circles corresponding to these paths , and we define @xmath317 so that @xmath318 takes each circle of @xmath316 to the closed based path that it corresponds to . we then subdivide @xmath316 so that @xmath318 is combinatorial . clearly the image of @xmath319 equals @xmath18 . observe that since @xmath316 is compact and @xmath320 is finite for each @xmath47-cell @xmath80 of @xmath6 , both @xmath321 and @xmath322 are finite . finally , note that @xmath316 is packed . beginning with @xmath316 , the algorithm produces a sequence of maps @xmath323 such that for each @xmath40 , @xmath324 is mapped onto @xmath18 . for each @xmath40 , @xmath325 is obtained from @xmath326 by either folding along a path in @xmath326 or by adding a copy of @xmath327 along a path @xmath73 in @xmath326 such that @xmath328 is a @xmath0-cell perimeter reduction . we will give a detailed description of these procedures below . each of these procedures will decrease the complexity and so we know that the sequence must terminate at a @xmath47-immersion @xmath329 such that @xmath330 does not admit a @xmath0-cell perimeter reduction . since @xmath6 satisfies the @xmath0-cell reduction hypothesis , we conclude that @xmath331 induces a @xmath7-injection , and therefore maps @xmath332 isomorphically onto @xmath18 , thus yielding a finite presentation for @xmath18 . as will be seen from the descriptions given below , each of these procedures can be implemented algorithmically . assume inductively that @xmath326 is compact and packed and that @xmath333 maps @xmath334 onto @xmath18 . * folding along a path : * if @xmath335 is not an immersion on @xmath336 , then there exists a map @xmath337 which is obtained by folding along a path . there is also a map @xmath338 such that @xmath335 factors as @xmath339 . because @xmath340 is @xmath7-surjective , we see that @xmath341 maps @xmath342 onto @xmath18 . thus by lemma [ lem : surj ] we have @xmath343 . since we also have @xmath344 , we see that the complexity of @xmath341 is strictly less than the complexity of @xmath335 . we can continue folding along paths until we reach a map @xmath345 where the restriction of @xmath346 to @xmath347 is an immersion . at this point we begin looking for a @xmath0-cell perimeter reduction . * adding @xmath0-cells : * suppose that the restriction of @xmath335 to @xmath336 is an immersion , but that there exists a @xmath0-cell perimeter reduction , @xmath348 . the reduction can be chosen to be maximal , and the result is a @xmath0-cell reduction which is either complete or incomplete . in both cases we define @xmath325 to be the identification space @xmath349 obtained by identifying the @xmath47-cells of @xmath73 in @xmath239 with their images in @xmath326 under the map @xmath350 . the map @xmath351 is well defined since @xmath352 and @xmath353 agree on the respective images of the @xmath47-cells of @xmath73 which were identified to form @xmath325 . also it is easy to see that the natural map @xmath354 is @xmath7-surjective , and therefore since @xmath355 we conclude that the @xmath7-image of @xmath341 is @xmath18 . it is again clear that @xmath325 is packed and compact . finally , the complexity of @xmath341 is strictly less than the complexity of @xmath335 since by lemma [ lem : attachments ] @xmath356 . the sequence of spaces described above is very similar to the sequence of spaces which would be constructed by theorem [ thm : gct ] when @xmath6 satisfies the @xmath0-cell reduction hypothesis . the main difference between the two is that the sequence of spaces in the proof of theorem [ thm : gct ] are subcomplexes of the covering space @xmath202 . in an algorithmic approach , the structure of this covering space is unavailable and the spaces described above have been constructed without reference to the space @xmath8 . in fact , these spaces may not embed or even immerse into @xmath8 throughout the course of the proof . we also note the following features of the algorithm : 1 . the algorithm gives an alternate proof of theorem [ thm:2cellct ] . the compactness assumption can be replaced by an appropriate recursiveness hypothesis and the algorithm is still effective . the algorithm can be used to prove coherence when the hypothesis is weakened to allow for weak @xmath0-cell perimeter reductions . in this more general context , however , one will not know when to stop running the algorithm and the algorithm as stated can not be used effectively , even when @xmath6 is compact . we note , however , that oliver payne @xcite has developed a variation of our algorithm which is effective for weak @xmath0-cell perimeter reductions so long as all of the sides have positive weights . we conclude this section with an estimate of the efficiency of the algorithm for finding finite presentations . [ cor : quadratic - time ] let @xmath6 be a compact weighted @xmath0-complex which satisfies the perimeter reduction hypothesis and let @xmath316 be the based bouquet of @xmath82 circles corresponding to a set of generators of a subgroup of @xmath14 . there exist constants @xmath357 and @xmath358 depending only on @xmath6 such that the algorithm described above terminates in fewer than @xmath359 steps and the time it takes to complete each step is bounded by @xmath360 , where @xmath361 denotes the number of @xmath47-cells in @xmath316 . in particular , the algorithm to calculate a finite presentation for the subgroup with these @xmath82 generators is @xmath362 . since @xmath6 is compact , there is a bound @xmath13 on the perimeter of any @xmath47-cell in @xmath6 and a bound @xmath363 on the length of the boundary of a @xmath0-cell in @xmath6 . next , notice that both folds and perimeter reductions will decrease the integer @xmath364 . this is because a fold will decrease the number of @xmath47-cells without increasing the perimeter , while a perimeter reduction will decrease @xmath365 by 1 while the number of @xmath47-cells is increased by at most @xmath366 . since the perimeter remains nonnegative , the number of steps will be bounded by @xmath367 thus we can choose @xmath368 . notice that since the number of steps is @xmath369 and since each step adds at most a bounded number of @xmath47-cells , the number of @xmath47-cells in @xmath326 is also @xmath369 . and since @xmath336 is connected , the number of @xmath21-cells in @xmath326 is also @xmath369 . next we show that the time it takes to complete each step is @xmath369 . let @xmath370 be the number of @xmath47-cells in @xmath6 and let the @xmath47-skeleton of @xmath326 be represented as an adjacency list . to check for the existence of a fold in @xmath326 only requires an examination of the links ( the adjacency lists ) of each @xmath21-cell . in each link we only need to check @xmath371 @xmath47-cells before we either find a fold or exhaust the link . thus a single link can be checked in constant time . since it is well - known that the time it takes to implement a breadth - first search of a connected graph represented by adjacency lists is @xmath372 ( see ( * ? ? ? * section 23.2 ) ) , the time it takes to visit each @xmath21-cell in @xmath326 is @xmath369 , and thus checking for a fold in @xmath326 is @xmath369 . next , suppose that @xmath336 is immersed into @xmath6 . since @xmath6 is compact , there is a finite list of paths @xmath373 which can lead to @xmath0-cell perimeter reductions . given one of these paths and a @xmath21-cell @xmath374 in @xmath326 it takes a finite amount of time to check whether there is a lift of @xmath73 which starts at @xmath374 . the constant nature of this search depends on the fact that the @xmath47-skeleton of @xmath326 is immersed into @xmath6 . this guarantees that the links of the @xmath21-cells are bounded in size and that at each point there is at most one extension of the lift which is a viable candidate . thus the search for a @xmath0-cell perimeter reduction in this type of complex is also @xmath369 . since the final complex can easily be converted into a finite presentation in quadratic time , the proof is complete . in this section we provide a second , more technical application of theorem [ thm : gct ] . our new hypothesis will imply that any immersion which is not @xmath7-injective admits a @xmath0-cell attachment which does not increase the perimeter , but with additional restrictions . the new hypothesis will involve sequences of closed paths in the space @xmath6 . we begin with an example which shows why these technicalities might be desirable . [ exmp : infinite ] let @xmath6 be the standard @xmath0-complex of @xmath375 , and give @xmath6 the unit weighting . let @xmath376 , and let @xmath8 be the based cover of @xmath6 corresponding to @xmath18 . as illustrated in figure [ fig : infinite ] , @xmath8 is an infinite cylinder . observe that every proper @xmath7-surjective subcomplex of @xmath202 admits a weak @xmath0-cell perimeter reduction . this example will show that hypothesizing weak @xmath0-cell perimeter reductions is insufficient to guarantee that the process of successively attaching @xmath0-cells will stop . consider the situation where we begin with the subcomplex @xmath4 which is the image of the closed path @xmath377 in @xmath8 . this subcomplex is shown on the left side of figure [ fig : infinite ] . the four @xmath47-cells determine a length @xmath378 closed path which is the boundary path of a @xmath0-cell in the cylinder . note that the inclusion map @xmath379 is not @xmath7-injective . although there is an obvious complete @xmath0-cell attachment which will make the inclusion a @xmath7-injection , it is also possible to apply an infinite sequence of @xmath0-cell attachments which are _ weak _ @xmath0-cell perimeter reductions , but at each stage the inclusion map will still fail to be @xmath7-injective . these @xmath0-cell attachments are formed by adding squares above or below the square hole bounded by the original closed path . the right side of figure [ fig : infinite ] shows the subcomplex obtained by adding two squares above the original closed path and two squares below it . the perimeter is @xmath380 , which is the same as @xmath5 . clearly , the operation of adding squares which do not change the perimeter can continue indefinitely . we conclude that a weak version of the @xmath0-cell reduction hypothesis , in and of itself , is insufficient to guarantee that a @xmath7-injective subcomplex will be obtained after a finite number of steps . the reason why we never reach a @xmath7-injective subcomplex in example [ exmp : infinite ] is that attached @xmath0-cells were not linked in any way to the failure of the @xmath7-injectivity . our plan will be that the order in which the @xmath0-cell attachments are applied will be tied to the existence of curves which are essential in @xmath4 and null - homotopic in @xmath8 . such precision was not needed for the @xmath0-cell reduction hypothesis because the process automatically stopped after a finite number of steps at a @xmath7-injective subcomplex . suppose that the path @xmath215 is obtained from @xmath71 by pushing across the @xmath0-cell @xmath51 . according to definition [ def : pushing - across ] , this means there is a certain subpath @xmath73 of @xmath69 which is also a subpath of @xmath87 , and @xmath220 is obtained from @xmath69 by replacing @xmath73 with its complement @xmath159 in @xmath50 . we will now augment this definition with certain perimeter requirements . if , in addition , @xmath381 , then @xmath220 is obtained from @xmath69 by _ pushing across a perimeter - reducing @xmath0-cell_. similarly , if @xmath266 , then it is obtained by _ pushing across a weakly perimeter - reducing @xmath0-cell_. [ def : path - reductions ] we say that a weighted @xmath0-complex @xmath6 satisfies the _ path reduction hypothesis _ provided the following condition holds : for every nontrivial closed null - homotopic path @xmath213 , there is a sequence of closed paths @xmath382 which starts at the path @xmath383 , ends at trivial path @xmath384 , and for each @xmath40 , @xmath385 is obtained from @xmath386 by either the removal of a backtrack or a weakly perimeter - reducing push across a @xmath0-cell . two elementary conditions which imply the path reduction hypothesis are a decrease in length and a descrease in area . in order to make the second condition precise we recall the defintion of area of a disc diagram . [ def : diagrams ] a _ disc diagram _ @xmath387 is a compact contractible @xmath0-complex with a fixed embedding in the plane . boundary cycle _ @xmath69 of @xmath387 is a closed path in @xmath388 which travels entirely around @xmath387 ( in a manner respecting the planar embedding of @xmath387 ) . let @xmath68 be a closed null - homotopic path . a _ disc diagram in @xmath6 for @xmath69 _ is a disc diagram @xmath387 together with a map @xmath389 such that the closed path @xmath68 factors as @xmath390 where @xmath391 is the boundary cycle of @xmath387 . the van kampen s lemma @xcite essentially states that every null - homotopic path @xmath68 is the boundary cycle of a disc diagram . we define @xmath392 to be the number of @xmath0-cells in @xmath387 . for a null - homotopic path @xmath68 , we define @xmath393 to equal the minimal number of @xmath0-cells in a disc diagram @xmath389 that has boundary cycle @xmath69 . the disc diagram @xmath389 will then be referred to as a _ minimal area disc diagram _ for @xmath69 . [ lem : path - reductions ] each of the following implies the path reduction hypothesis : 1 . every immersed nontrivial null - homotopic path @xmath68 admits a push across a weakly perimeter - reducing @xmath0-cell which yields a strictly shorter path @xmath394 . every immersed nontrivial null - homotopic path @xmath68 admits a push across a weakly perimeter - reducing @xmath0-cell which yields a path @xmath394 satisfying @xmath395 . in either case , there is an obvious procedure for creating the sequence of paths @xmath396 which starts at a given closed null - homotopic path @xmath213 and ends at the trivial path . we first remove backtracks repeatedly until we obtain an immersed path , then use the condition to find a weakly perimeter - reducing push across a @xmath0-cell , and then repeat . in each case , the process must terminate at a trivial path after finitely many steps because the removal of backtracks does not increase either length or area . we will now show that the path reduction hypothesis implies coherence . [ thm : pathct ] if @xmath6 is a weighted @xmath0-complex which satisfies the path reduction hypothesis , then @xmath6 satisfies the perimeter reduction hypothesis , and thus @xmath14 is coherent . we will assume that the map @xmath3 is packed , for otherwise we could attach @xmath0-cells to form a packed map @xmath297 with @xmath397 and with the inclusion @xmath398 a @xmath7-surjection . if @xmath140 is not @xmath7-injective then there is a closed essential path @xmath311 such that the composition @xmath399 is a null - homotopic path in @xmath6 , and by the path reduction hypothesis there exists a sequence of paths @xmath400 for @xmath401 which starts at @xmath402 , ends at the trivial path @xmath403 , and for each @xmath40 , @xmath385 is obtained from @xmath386 by either removing a backtrack or a weakly perimeter - reducing push across a @xmath0-cell . we will use this sequence of paths to create a sequence of compact spaces @xmath404 and maps @xmath405 and a sequence of paths @xmath406 which are lifts of the paths @xmath396 . let @xmath407 and let @xmath408 equal @xmath311 and assume that @xmath326 and @xmath406 have been defined for some @xmath40 . the space @xmath325 is obtained from @xmath326 as follows . if the operation transforming @xmath386 into @xmath385 is either the removal of a backtrack or a weakly perimeter - reducing push across a @xmath0-cell @xmath50 where the map @xmath240 already lifts to @xmath4 at the appropriate point , then @xmath409 . the exact requirement in the latter case is that @xmath240 lift to a map @xmath410 such that the composition @xmath411 is the map @xmath412 obtained by restricting the path @xmath413 . since @xmath414 it is clear that @xmath415 . the path @xmath416 is defined to be the obvious modification of the path @xmath406 . if the operation is a weakly perimeter - reducing push across a @xmath0-cell @xmath50 and the map @xmath240 does not lift to @xmath4 at the appropriate point , then @xmath325 is defined to be @xmath417 and @xmath385 is again the obvious modification of the path @xmath418 . figure [ fig : reductionsite ] illustrates a @xmath0-cell attachment which arises in this way . the technical condition is that there does not exist a lift to a map @xmath410 such that the composition @xmath419 is the map @xmath420 obtained by restricting the path @xmath421 . notice that in this case @xmath422 is a @xmath0-cell attachment which is a weak @xmath0-cell perimeter reduction . if this attachment is complete , then by lemma [ lem : complete ] , @xmath356 . if the attachment is incomplete , then by lemma [ lem : incomplete ] , @xmath423 . in each instance the path @xmath416 is obtained from @xmath424 by lifting the operation which occurred in @xmath6 to @xmath325 . combining the sequence of perimeter inequalities we see that @xmath425 with a strict inequality if any of the @xmath0-cell attachments were complete attachments . it only remains to show that at least one of the attachments was complete . notice that at each stage the closed path @xmath426 is homotopic to the closed path @xmath427 . the crucial observation is that the removal of a backtrack or an incomplete attachment can never change an essential path into a null - homotopic one . since the final path @xmath384 is null - homotopic , at least one of the attachments must have been complete . thus @xmath428 . to complete the proof , let @xmath429 . we note @xmath430 is @xmath7-surjective because it is the composition @xmath431 and for each @xmath40 the map @xmath432 is either a homeomorphism , a complete attachment or an incomplete attachment and thus always a @xmath7-surjection . [ rem : subspace sequence ] let @xmath6 be a weighted @xmath0-complex which satisfies the path reduction hypothesis . if @xmath4 is a compact connected subcomplex of a cover @xmath8 of @xmath6 , then according to theorem [ thm : gct ] and theorem [ thm : pathct ] , there exists a sequence of subcomplexes @xmath433 such that for @xmath434 , the space @xmath325 is the image of @xmath435 in @xmath202 , where @xmath436 is a @xmath0-cell attachment which is a weak perimeter reduction and @xmath66 is a @xmath7-isomorphic subcomplex of @xmath8 . theorem [ thm : pathct ] and remark [ rem : subspace sequence ] lead to a pair of interesting corollaries . the first corollary is a bound on the number of relators needed in the presentation for a finitely generated subgroup . [ cor : relatorbound ] let @xmath6 be a weighted @xmath0-complex which satisfies the path reduction hypothesis and assume that each @xmath0-cell of @xmath6 is attached along a simple cycle in @xmath437 . if @xmath438 is a finite set of closed based paths in @xmath6 , then the finitely generated subgroup @xmath439 generated by the closed based paths @xmath440 has a finite presentation with at most @xmath441 relations . similarly , for any @xmath7-surjective subcomplex @xmath204 whose perimeter is finite and whose fundamental group is free , there is a finite presentation for @xmath442 where the number of relators is at most @xmath443 . we first note that the second assertion includes the first assertion as a special case . in particular , given closed paths @xmath444 we can take the based cover @xmath445 which corresponds to @xmath18 and lift the closed paths @xmath446 to closed paths @xmath447 with a common basepoint . the union of the images of these closed paths is a finite graph @xmath4 in @xmath445 which satisfies the description in the second half of the corollary . thus it suffices to prove the second assertion . according to remark [ rem : subspace sequence ] , there is a sequence of subspaces @xmath448 such that for @xmath449 , the space @xmath325 is the image of @xmath435 in @xmath8 , where @xmath450 is a @xmath0-cell attachment which is a weak perimeter reduction , and such that @xmath66 is a @xmath7-isomorphic subcomplex of @xmath8 . if all of the @xmath0-cell attachments are perimeter reductions , then the argument is easy , because then @xmath451 and so @xmath66 can be obtained from @xmath4 by the addition of fewer than @xmath5 @xmath0-cells . consequently , @xmath192 has a presentation with fewer than @xmath5 @xmath0-cells and we are done . when some of the @xmath0-cell attachments are weak perimeter reductions , we argue as follows : let @xmath452 be the minimum number of relators which are needed to give a finite presentation of @xmath334 . we will show inductively that @xmath453 . this is true for @xmath454 since the fundamental group of @xmath455 is free . suppose that @xmath325 is obtained from @xmath326 by an incomplete attachment . in this case there is a new @xmath47-cell which is added to @xmath325 and this @xmath47-cell appears exactly once in the attaching map of the new @xmath0-cell . this is where we use the additional hypothesis that the attaching map of each @xmath0-cell embeds in @xmath6 . now we can collapse the new @xmath0-cell across this new @xmath47-cell to see that no new relations have been added , although this @xmath0-cell attachment may have added new generators . since the perimeter has not increased , the inequality is still true . if , on the other hand , @xmath325 is obtained from @xmath326 by a complete attachment , then @xmath456 , but @xmath457 . thus the inequality holds in this case . when the process stops , the perimeter is still nonnegative and thus @xmath458 is bounded by @xmath321 , which is the assertion . note that a similar result ( using essentially the same proof ) can be proved under the assumption that each @xmath0-cell is attached along a ( possibly trivial ) power of some simple cycle . the same type of proof can also be used to provide an upper bound on the euler characteristic of a subgroup . [ cor : eulerbound ] let @xmath6 be a weighted @xmath0-complex which satisfies the path reduction hypothesis and assume that no @xmath0-cell of @xmath6 is attached by a proper power . for any @xmath7-surjective compact subcomplex @xmath459 there is a compact @xmath7-isomorphic subcomplex @xmath460 such that @xmath461 . according to remark [ rem : subspace sequence ] , there is a sequence of subspaces @xmath448 such that for @xmath449 , the space @xmath325 is the image of @xmath435 in @xmath8 , where @xmath450 is a @xmath0-cell attachment which is a weak perimeter reduction , and such that @xmath66 is a @xmath7-isomorphic subcomplex of @xmath8 . we will deduce that @xmath461 by showing that for each @xmath40 we have @xmath462 . for each @xmath40 , @xmath325 is the union of @xmath326 and the closure of some @xmath0-cell . first suppose that @xmath463 . in this case , @xmath325 is obtained from @xmath326 by the addition of a single @xmath0-cell and so @xmath464 but @xmath465 , so the inequality holds . next suppose that @xmath466 , in which case @xmath467 because while a @xmath0-cell has been added , at least one nontrivial arc of @xmath47-cells is added to @xmath468 along its endpoints . since @xmath469 , we see that @xmath470 . we note that a similar statement can be proved in case some of the @xmath0-cells are attached by proper powers . we close the section with the following problem . [ prob : coherence ] let @xmath6 be a compact weighted @xmath0-complex . suppose that @xmath471 for every minimal area disc diagram @xmath472 . does it follow that @xmath14 is coherent ? we conjecture that the answer is yes , but it is not clear how to proceed . the problem which arises is that maps @xmath473 which do not send the sides of the boundary of @xmath474 injectively to the sides of @xmath6 are , in a fairly strong sense , unavoidable . in this section we present a criterion for the coherence of one - relator groups with torsion , followed by some applications . additional criteria for the coherence of other types of one - relator groups are developed in @xcite and a similar criterion will be described for small cancellation groups in section [ sec:2-cell - small ] . the coherence criterion for one - relator groups is a combination of theorem [ thm : pathct ] and the `` spelling theorem '' of b.b . ( the original reference is @xcite ; see @xcite and @xcite for combinatorial and geometric proofs . ) here is the theorem as it is usually formulated . [ thm : spelling ] let @xmath475 where @xmath23 is a cyclically reduced word and @xmath476 . let @xmath477 and @xmath478 be words in @xmath479 which are equivalent in @xmath19 . if @xmath477 is freely reduced and @xmath478 omits a generator which occurs in @xmath477 , then @xmath477 contains a subword @xmath480 which is also a subword of @xmath229 and latexmath:[$|w'| > which represents the identity in @xmath19 , then @xmath477 contains such a subword @xmath480 . because of the correspondence between presentations and their standard @xmath0-complexes , we will express our main theorem about one - relator groups in the language of @xmath0-complexes . recall that by convention [ conv : combinatorial ] the @xmath0-complexes under consideration will be those which correspond to presentations whose defining relators are cyclically reduced . [ thm : wt - one - rel ] let @xmath6 be a weighted @xmath0-complex with a unique @xmath0-cell @xmath241 and a unique @xmath21-cell . let @xmath482 be the period and let @xmath476 be the exponent of @xmath54 . if the inequality @xmath273 holds for every subpath @xmath159 of @xmath87 satisfying @xmath483 , then @xmath14 is coherent . let @xmath69 be a closed immersed null - homotopic path in @xmath6 and let @xmath477 be the word corresponding to @xmath69 in the generators of the presentation corresponding to @xmath6 . since @xmath69 is immersed , @xmath477 is freely reduced , and so by theorem [ thm : spelling ] there exists a subpath @xmath73 in @xmath69 such that @xmath73 is a subpath of @xmath87 and @xmath484 . note that we are applying the spelling theorem in the special case where @xmath478 is the trivial word . since the complement of @xmath73 is a path @xmath485 with @xmath483 , we know by assumption that @xmath273 . by lemma [ lem : s - form ] , it follows that @xmath486 . therefore @xmath69 can be pushed across a weakly perimeter - reducing @xmath0-cell . moreover , the new path obtained by replacing @xmath73 with @xmath159 , is strictly shorter than @xmath69 because @xmath487 . thus by lemma [ lem : path - reductions ] , @xmath6 satisfies the path reduction hypothesis , and so @xmath14 is coherent by theorem [ thm : pathct ] . as an application of theorem [ thm : wt - one - rel ] , we obtain the following : [ thm : equalweights ] let @xmath23 be a cyclically reduced word and let @xmath24 . if @xmath25 , then @xmath19 is coherent . in particular , for every word @xmath23 , the group @xmath488 is coherent provided that @xmath26 is sufficiently large . let @xmath6 be the standard @xmath0-complex of the presentation with the unit weighting . we can assume that @xmath476 since otherwise @xmath19 is virtually free , and hence obviously coherent . without loss of generality we can also assume that @xmath23 is not a proper power since this would only serve to make the hypothesis more stringent . let @xmath50 denote the unique @xmath0-cell of @xmath6 , and regard the word @xmath23 as a path @xmath482 . then @xmath54 has period @xmath482 and has exponent @xmath26 . since the perimeter of a @xmath47-cell @xmath55 in @xmath6 will be the number of times its associated generator occurs in @xmath489 in either orientation , we can estimate that @xmath490 , and that @xmath491 for any word @xmath492 with @xmath483 . on the other hand , the weight of the @xmath0-cell @xmath50 is exactly @xmath493 . the coherence criterion of theorem [ thm : wt - one - rel ] will be satisfied so long as @xmath494 in particular , if @xmath495 then theorem [ thm : wt - one - rel ] shows that the group @xmath19 is coherent . the next theorem lowers the bound on the exponent by choosing a more appropriate weight function . [ thm : mingenerator ] let @xmath23 be a cyclically reduced word , let @xmath496 , and let @xmath497 occurs exactly @xmath498 times @xmath499 in the word @xmath23 . if @xmath500 , then the group @xmath19 is coherent . in particular , if every @xmath43 ( @xmath501 ) occurs in @xmath23 , then @xmath19 is coherent for all @xmath502 . the proof is nearly identical to the previous one , except that the weight function on @xmath6 has changed . assign a weight of @xmath47 to any side labeled by the generator @xmath497 and assign a weight of @xmath21 otherwise . the perimeter of the @xmath47-cell labeled by @xmath497 is exactly @xmath503 ( since this is the number of occurrences of @xmath497 in @xmath229 ) , and the perimeter of any other @xmath47-cell is @xmath21 . since any word @xmath504 with @xmath505 contains at most @xmath498 @xmath47-cells labeled by @xmath497 , we estimate that @xmath506 . on the other hand , @xmath305 is exactly @xmath507 . thus whenever @xmath500 , the criterion of theorem [ thm : wt - one - rel ] will be satisfied , and the group @xmath508 will be coherent . the final assertion is immediate since the word @xmath23 contains at least @xmath82 letters and thus one of them occurs at most @xmath509 times . in this section we apply our coherence results to small cancellation groups . we begin with a brief review of the basic notions of small cancellation theory . the reader is referred to @xcite for a rigorous development of these notions that is consistent with their use here . [ def : piece ] let @xmath6 be a combinatorial @xmath0-complex . intuitively , a piece of @xmath6 is a path which is contained in the boundaries of the @xmath0-cells of @xmath6 in at least two distinct ways . more precisely , a nontrivial path @xmath68 is a _ piece _ of @xmath6 if there are @xmath0-cells @xmath162 and @xmath163 such that @xmath68 factors as @xmath510 and as @xmath511 but there does not exist a homeomorphism @xmath278 such that there is a commutative diagram @xmath512 excluding commutative diagrams of this form ensures that @xmath69 occurs in @xmath513 and @xmath514 in essentially distinct ways . [ def : sc ] an _ arc _ in a diagram is a path whose internal vertices have valence @xmath0 and whose initial and terminal vertices have valence @xmath515 the arc is _ internal _ if it its interior lies in the interior of @xmath387 , and it is a _ boundary arc _ if it lies entirely in @xmath388 . a @xmath0-complex @xmath6 satisfies the @xmath516 condition if for every minimal area disc diagram @xmath389 , each internal @xmath21-cell of @xmath387 has valence @xmath0 or valence @xmath517 . similarly , @xmath6 satisfies the @xmath518 condition if the boundary path of each @xmath0-cell in @xmath387 either contains a nontrivial boundary arc , or is the concatenation of at least @xmath519 nontrivial internal arcs . finally , for a fixed positive real number @xmath520 , the complex @xmath6 satisfies @xmath521 provided that for each @xmath0-cell @xmath51 , and each piece @xmath68 which factors as @xmath522 , we have @xmath523 . note that if @xmath6 satisfies @xmath521 and @xmath524 then @xmath6 satisfies @xmath525 . it is a fact that if @xmath389 is minimal area then each nontrivial arc in the interior of @xmath387 is a piece in the sense of definition [ def : piece ] . although the rough definition given above is not quite technically correct ( for instance , it uses minimal area diagrams instead of reduced diagrams ) , it should give the reader unfamiliar with small cancellation complexes an approximate idea of their properties . we refer the interested reader to @xcite for precise definitions . when @xmath519 and @xmath526 are sufficiently large , minimal area diagrams over @xmath6 will always contain either spurs or @xmath40-shells . [ def : i - shells ] let @xmath387 be a diagram . an _ @xmath40-shell _ of @xmath387 is a @xmath0-cell @xmath527 whose boundary cycle @xmath87 is the concatenation @xmath528 where @xmath529 is a boundary arc , the interior of @xmath530 maps to the interior of @xmath387 , and @xmath531 is a nontrivial interior arc of @xmath387 for all @xmath532 . the path @xmath533 is the _ outer path _ of the @xmath40-shell . a @xmath47-cell @xmath55 in @xmath388 which is incident with a valence @xmath47 @xmath21-cell @xmath374 is a _ spur_. in analogy with the outer path of an @xmath40-shell , we will refer to the length @xmath0 path ( either @xmath214 or @xmath534 ) that passes through @xmath374 as the _ outer path _ of the spur . illustrated from left to right in figure [ fig : outercells ] are disc diagrams containing a spur , a @xmath21-shell , a @xmath47-shell , a @xmath0-shell , and a @xmath1-shell . in each case , the @xmath0-cell @xmath50 is shaded , and the boundary arc @xmath533 is @xmath535 . -shells [ fig : outercells ] ] the classical result which forms the basis of small cancellation theory is called greendlinger s lemma ( see ( * ? ? ? * thm v.4.5 ) ) . the following strengthening of greendlinger s lemma was proven in ( * ? ? ? * thm 9.4 ) . while the results of this section only require greendlinger s lemma itself , we will require the full strength of the following theorem in section [ sec : sc - lqc ] . [ thm : fan - classification ] if @xmath387 is a @xmath536-@xmath537 @xmath538-@xmath539 $ ] disc diagram , then one of the following holds : 1 . @xmath387 contains at least three spurs and/or @xmath40-shells with @xmath540 @xmath541 $ ] . 2 . @xmath387 is a ladder of width @xmath542 , and hence has a spur , @xmath21-shell or @xmath47-shell at each end . @xmath387 consists of a single @xmath21-cell or a single @xmath0-cell . moreover , if @xmath387 is nontrivial and @xmath374 is a @xmath21-cell in @xmath388 , then @xmath387 contains a spur or an @xmath40-shell with @xmath543 @xmath544@xmath545@xmath546 $ ] which avoids @xmath374 , and if the cut - tree of @xmath387 has @xmath547 leaves , then @xmath387 contains at least @xmath547 separate such spurs and @xmath40-shells . see @xcite for details . in the present article we will only need the following immediate corollary . [ cor : fan - classification ] let @xmath387 be a @xmath536-@xmath537 @xmath538-@xmath539 $ ] disc diagram and let @xmath69 and @xmath73 be immersed paths such that @xmath548 is the boundary cycle of @xmath387 . if neither path contains the outer path of an @xmath40-shell in @xmath387 with @xmath543 @xmath541 $ ] , then every @xmath0-cell of @xmath387 contains an edge in @xmath69 and an edge in @xmath73 . [ thm : coh - sc-1 ] let @xmath6 be a weighted @xmath0-complex which satisfies @xmath549-@xmath550 @xmath551-@xmath552 $ ] . suppose @xmath553 for each @xmath0-cell @xmath113 and path @xmath485 which is the concatenation at most three @xmath544two@xmath546 $ ] consecutive pieces in the boundary of @xmath50 . then @xmath14 is coherent . we will prove the @xmath549-@xmath550 case ; the @xmath536-@xmath537 case is handled similarly . by theorem [ thm : pathct ] , it is sufficient to show that @xmath6 satisfies the path reduction hypothesis . let @xmath212 be a closed immersed nontrivial null - homotopic path . let @xmath389 be a minimal area disc diagram with boundary cycle @xmath69 . according to theorem [ thm : fan - classification ] , there exists an @xmath40-shell of @xmath387 ( @xmath554 ) which avoids the basepoint of @xmath69 . by hypothesis , the new boundary path , obtained by removing the boundary arc and the @xmath0-cell of this @xmath40-shell from the diagram , is a path @xmath220 which can be obtained from @xmath69 by a weakly perimeter - reducing push across a @xmath0-cell . since @xmath220 is a path satisfying @xmath555 , lemma [ lem : path - reductions ] shows that @xmath6 satisfies the path reduction hypothesis and the proof is complete . theorem [ thm : coh - sc-1 ] can be improved by using more complicated weight functions , by using more complicated reductions , or by altering the presentation substantially before a weight function is applied . the following example is an illustration of the latter possibility . additional examples can be found in section [ sec : sc - lqc ] and in @xcite . [ exmp : modify ] consider the following one - relator group . @xmath556 since the relator is not a proper power , the theorems in section [ sec:1relator ] do not apply . if we alter the presentation of @xmath19 by introducing a new generator @xmath557 , then @xmath558 and the new presentation satisfies certain small cancellation conditions . this can be seen from the link of the @xmath21-cell of the standard @xmath0-complex @xmath6 of the modified presentation . as illustrated in figure [ fig : modify ] , the link is simplicial and so all pieces are of length @xmath47 , and since both relators have length at least @xmath559 , @xmath6 is a @xmath549 presentation . because the shortest circuit in the link has length @xmath378 , @xmath6 satisfies @xmath537 . if we assign a weight of @xmath47 to each of the sides in the relations which are labeled by @xmath560 or @xmath55 , and assign a weight of @xmath21 to the sides labeled @xmath561 , then the corresponding @xmath47-cell perimeters and @xmath0-cell weights are as follows . the @xmath47-cells labeled @xmath560 , and @xmath55 have a perimeter of @xmath0 , and both @xmath0-cells have a weight of @xmath562 . since the presentation satisfies @xmath549-@xmath537 we can use the coherence criterion for @xmath536-@xmath537-complexes ( theorem [ thm : coh - sc-1 ] ) . the criterion is satisfied since @xmath563 and @xmath564 for all appropriate @xmath50 and @xmath73 . consequently this group is coherent . -cell of the standard @xmath0-complex for the presentation in example [ exmp : modify ] [ fig : modify ] ] the reader may have noticed that although a different weight is allowed for each side of each @xmath0-cell of @xmath6 , in all of the examples we have given so far , we have always chosen the weights to be equal on all of the sides incident at any particular @xmath47-cell in @xmath6 . since it is clear that the perimeter of a @xmath47-cell @xmath55 in @xmath6 is unaffected by the distribution of the weights among the sides present at @xmath55 so long as their total is left invariant , this raises the question of whether the added flexibility we have allowed will ever be needed . in our final example we show that the weights of the sides at @xmath55 sometimes do need to be different . consider a presentation of the form @xmath565 . suppose that for each @xmath40 , the generator @xmath43 appears exactly the same number of times in @xmath477 as in @xmath478 , so that in particular @xmath566 . and suppose further that the pieces of @xmath478 are longer than the pieces of @xmath477 . this is the situation in which it makes sense that a side at @xmath43 in @xmath478 will need more weight than a side at @xmath43 in @xmath477 . the following is a concrete example . consider the two - relator presentation : @xmath567 the parentheses are included for emphasis only . we will call the first relator @xmath477 and the second relator @xmath478 . observe that the presentation is invariant under a cyclic shift of the generators . notice also that the presentation satisfies @xmath537 and @xmath568 , that every piece in @xmath477 has length @xmath47 , and that @xmath478 has pieces of length at most @xmath1 . finally it is clear that the subpath @xmath569 in @xmath478 is a union of two consecutive pieces . if we assign the sides of @xmath477 weight @xmath47 and we assign the sides of @xmath478 weight @xmath1 then the perimeter of each @xmath47-cell is @xmath570 and the weights of the @xmath0-cells corresponding to @xmath477 and @xmath478 are @xmath571 and @xmath572 respectively . consequently the coherence criterion of theorem [ thm : coh - sc-1 ] is satisfied and so the group is coherent . on the other hand , if we used the unit weighting , then the perimeter of each @xmath47-cell is @xmath380 . observe that the path @xmath569 has perimeter @xmath573 which is greater than the weight @xmath571 of the @xmath0-cell corresponding to @xmath478 , and so the criterion of theorem [ thm : coh - sc-1 ] fails . we will now show that more is true . for this presentation , there does not exist a way to assign weights to the sides of the @xmath0-cells so that ( 1 ) all of the sides labeled by a given generator receive the same weight , and ( 2 ) the coherence criterion of theorem [ thm : coh - sc-1 ] is satisfied . a set of weights which satisfies the coherence criterion of theorem [ thm : coh - sc-1 ] will be called _ satisfactory_. the argument now goes as follows : observe that the sum of any two sets of weights which are satisfactory will also be satisfactory , and that a cyclic shift of a set of weights which are satisfactory will remain satisfactory . next suppose that a set of weights existed which satisfied conditions ( 1 ) and ( 2 ) . by the above observations we could add this set of weights to all of its cyclic shifts to show that a scalar multiple of the unit perimeter is satisfactory . but since we know that the unit perimeter fails the weight criterion , this contradiction shows that no such set of weights can exist . in this section we introduce our final coherence hypotheses and our final coherence theorems which employ fans instead of single @xmath0-cells . many of the definitions , statements , and proofs will be analogous to those in previous sections . on the left is a fan @xmath574 whose outer path @xmath73 is the bold path on its boundary . the disc diagram on the right contains the fan @xmath574 as a subcomplex . note that @xmath73 is a subpath of the boundary path of @xmath387.,title="fig : " ] on the left is a fan @xmath574 whose outer path @xmath73 is the bold path on its boundary . the disc diagram on the right contains the fan @xmath574 as a subcomplex . note that @xmath73 is a subpath of the boundary path of @xmath387.,title="fig : " ] a _ fan _ @xmath574 is a @xmath0-complex homeomorphic to a closed disc , which is the union of closed @xmath0-cells @xmath575 , with the property that for each @xmath40 , @xmath576 is the disjoint union of the connected sets @xmath577 and @xmath578 ( note that when @xmath454 or @xmath579 one of these sets is empty . ) the _ outer path _ @xmath73 of @xmath574 is a concatenation @xmath580 where each @xmath581 is a subpath of @xmath582 . we refer the reader to figure [ fig : typicalfan ] for a picture of a typical fan . the unique path @xmath159 such that @xmath217 is the boundary cycle of @xmath574 will be called the _ inner path _ of @xmath574 . given a map @xmath583 there is a unique extension to a packed map ( definition [ def : packets ] ) where the @xmath47-skeleton of the domain is unchanged . we will denote this extended domain by @xmath584 in analogy with @xmath585 . we will only be interested in fans equipped with a map @xmath586 such that @xmath587 is a near - immersion , and we will refer to such a mapped fan as a _ fan in @xmath6_. in this case , we will also regard the outer path @xmath588 of @xmath574 as a path @xmath72 . the disc diagram @xmath389 _ contains the fan @xmath586 _ , provided that @xmath583 factors as @xmath589 , where the outer path @xmath73 of @xmath574 maps to @xmath388 , and the inner path @xmath159 of @xmath574 is an internal path in @xmath387 . the simplest fans are @xmath40-shells ( definition [ def : i - shells ] ) . in this case @xmath87 is the concatenation @xmath217 where @xmath73 is the outer path of the @xmath40-shell , and @xmath159 is the concatenation of @xmath40-pieces in @xmath6 . we will often be interested in a collection @xmath590 of fans in a @xmath0-complex @xmath6 which satisfy additional properties . the next three definitions are technical conditions which will enable us to perform a perimeter calculation for fan attachments parallel to the calculation in lemma [ lem : attachments ] . [ def : reducing - fan ] let @xmath6 be a weighted @xmath0-complex . the fan @xmath583 is _ perimeter - reducing _ provided that the perimeter of @xmath591 is less than the perimeter of its outer path @xmath72 . in other words , @xmath592 . similarly , it is _ weakly perimeter - reducing _ if @xmath593 . let @xmath3 be a packed @xmath47-immersion , let @xmath583 be a fan , and let @xmath594 be a lift of the outer path of @xmath574 to @xmath4 . we say @xmath574 is _ missing in @xmath4 along @xmath594 _ provided that the lift of @xmath595 to the path @xmath594 does not extend to a lift of any @xmath0-cell @xmath50 of @xmath584 that intersects @xmath73 in a nontrivial path . specifically , if @xmath50 is a @xmath0-cell of @xmath584 and @xmath596 is a nontrivial path , then there should not exist a lift of @xmath597 to @xmath4 which extends the path @xmath251 . equivalently @xmath583 is missing in @xmath4 along @xmath594 provided that for each @xmath47-cell @xmath526 in @xmath73 , no side of a @xmath0-cell of @xmath6 at @xmath80 is present at both @xmath81 and @xmath561 , where @xmath561 , @xmath81 , and @xmath80 are the images of @xmath526 in @xmath584 , @xmath4 , and @xmath6 . [ def : spread ] a fan @xmath583 is _ spread - out _ provided that the sides of @xmath0-cells of @xmath584 along @xmath47-cells in the outer path @xmath588 project to distinct sides of @xmath0-cells along @xmath47-cells in @xmath6 . this condition is certainly satisfied when the outer path @xmath598 projects to a path @xmath599 which does not pass through any @xmath47-cell of @xmath6 more than once . for instance , @xmath583 is spread - out when @xmath600 is an embedding , and it is spread - out when @xmath72 is a ( possibly closed ) simple path . the following lemma calculates the perimeter of @xmath601 in terms of the perimeters of its constituents . [ lem : fan - attachments ] let @xmath6 be a weighted @xmath0-complex , let @xmath46 be a packed @xmath47-immersion with @xmath602 , and let @xmath594 be a lift of the outer path of a fan @xmath586 . if @xmath583 is spread - out and @xmath574 is missing along @xmath594 , then , letting @xmath603 , the perimeter of the induced map @xmath285 satisfies : @xmath604 thus , if @xmath605 is perimeter - reducing then @xmath606 and if @xmath583 is weakly perimeter - reducing then @xmath308 . the proof is similar to that of lemma [ lem : incomplete ] where it is obvious that a fan consisting of a single @xmath0-cell is spread - out . since the perimeter of @xmath46 is unaffected by the addition or removal of redundant @xmath0-cells from @xmath4 ( lemma [ lem : redundant ] ) , we may assume that @xmath4 has no redundancies . by lemma [ lem : no - redundant->immersion ] this means that we may assume that @xmath58 is an immersion . next , we show that the map @xmath297 is a near - immersion . by the definition of a fan , distinct @xmath47-cells of @xmath73 are sent to distinct @xmath47-cells in @xmath584 under the map @xmath607 and thus distinct @xmath47-cells of @xmath4 are mapped to distinct @xmath47-cells of @xmath196 under the map @xmath608 . this shows that the induced map @xmath609 is an injection . on the other hand , the map @xmath610 is a near - immersion by definition . combined with the fact that @xmath583 is spread - out , this shows that the induced map @xmath611 is also an injection . thus , if @xmath198 fails to be a near - immersion , it must fail along the path @xmath298 . more precisely , to show that @xmath198 is a near - immersion , it only remains to be shown that a side of @xmath6 which lifts to a side of @xmath73 in @xmath584 could not also lift to a side of @xmath73 in @xmath4 . this is impossible because of the assumption that @xmath574 is missing along @xmath594 . if we assume for the moment that @xmath4 is compact , then we can calculate the perimeter of @xmath196 using equation ( [ eq : weight - stuff ] ) of lemma [ lem : perimeter - rewritten ] . according to equation ( [ eq : weight - stuff ] ) , the perimeter of @xmath196 equals the sum of the perimeters of its @xmath47-cells minus the weights of its @xmath0-cells . if we apply equation ( [ eq : weight - stuff ] ) to @xmath584 and @xmath4 separately then we would add the perimeters of their @xmath47-skeletons and subtract the weights of their @xmath0-cells . the difference between these counts arises from the @xmath47-cells of @xmath73 in @xmath584 which get identified with @xmath47-cells of @xmath4 in the space @xmath196 . this proves equation ( [ eq : fan ] ) . in the general case where @xmath5 is finite but @xmath4 is not compact , the proof proceeds as in lemma [ lem : incomplete ] , except that lemma [ lem : perimeter - rewritten ] is applied to the packed fan @xmath584 instead of the packet @xmath585 . having established conditions under which we can control the change in perimeter , it is now relatively easy to define a hypothesis and prove a coherence theorem . [ def : fanrh ] a packed @xmath47-immersion @xmath46 admits a _ fan perimeter reduction _ provided there is a perimeter - reducing spread - out fan @xmath612 and there exists a lift of its outer path to @xmath4 such that @xmath574 is missing along @xmath251 . a weighted @xmath0-complex @xmath6 satisfies the _ fan reduction hypothesis _ if each packed @xmath47-immersion @xmath46 which is not @xmath7-injective , admits a fan perimeter reduction . [ thm : fanct ] let @xmath6 be a weighted @xmath0-complex . if @xmath6 satisfies the fan reduction hypothesis then @xmath6 satisfies the perimeter reduction hypothesis , and thus @xmath14 is coherent . let @xmath140 be a @xmath47-immersion which is not @xmath7-injective . since adding the @xmath0-cells necessary to make @xmath3 a packed map does not increase perimeter , we may assume it is packed without loss of generality . by hypothesis , there is a perimeter - reducing fan @xmath583 which is spread - out and a lift of its outer path to @xmath4 such that @xmath574 is missing along @xmath594 . this can be used to create a complex @xmath196 whose perimeter is smaller , by lemma [ lem : fan - attachments ] . the fact that @xmath4 and @xmath196 have the same @xmath7 image in @xmath6 is obvious . thus @xmath6 satisfies the perimeter reduction hypothesis . in most applications , we will only use a special case of theorem [ thm : fanct ] which can be formulated in terms of disc diagrams . [ thm : fan - diagramct ] let @xmath6 be a weighted @xmath0-complex and let @xmath590 be a collection of perimeter - reducing spread - out fans in @xmath6 . if each nontrivial minimal area disc diagram @xmath389 contains a spur or a fan from @xmath590 , then @xmath6 satisfies the perimeter reduction hypothesis , and thus @xmath14 is coherent . by theorem [ thm : fanct ] , it is sufficient to show that @xmath6 satisfies the fan reduction hypothesis . let @xmath4 be a compact @xmath0-complex and let @xmath140 be a packed @xmath47-immersion which is not @xmath7-injective . there exists at least one essential immersed closed path @xmath613 whose image in @xmath6 is a null - homotopic closed immersed path . we assume that @xmath613 has been chosen so that @xmath393 is as small as possible , and such that the disc diagram @xmath389 realizes this minimum area . by hypothesis , @xmath387 contains a fan @xmath614 which is perimeter - reducing and spread - out . let @xmath594 be the restriction of @xmath311 to the outer path of @xmath574 . since @xmath69 was chosen to have minimal area , the fan @xmath574 is missing along @xmath251 . consequently @xmath6 satisfies the fan reduction hypothesis . indeed , if some some side of a @xmath0-cell @xmath50 in @xmath584 is already present along some edge @xmath526 of the path @xmath69 , then @xmath69 is homotopic in @xmath4 to a path @xmath220 which travels around the boundary of @xmath615 . but @xmath616 , so we can find an immersed essential path in @xmath4 which bounds a smaller area diagram in @xmath6 , and this is impossible . we conclude this section with two further generalizations of results from the previous sections . [ thm : fan algorithm ] if @xmath6 is a compact weighted @xmath0-complex which satisfies the fan reduction hypothesis , then there is an algorithm which produces a finite presentation for any subgroup of @xmath14 given by a finite set of generators . the proof is analogous to the proof of theorem [ thm : algorithm ] and we leave the details to the reader . let @xmath6 be a compact weighted @xmath0-complex . let @xmath590 be a collection of spread - out weakly perimeter - reducing fans , and suppose that for each fan @xmath617 , we have @xmath618 . if every nontrivial minimal area disc diagram contains a spur or a fan in @xmath590 then @xmath14 is coherent . the proof is essentially a generalization of the proof of theorem [ thm : pathct ] that uses fans instead of @xmath0-cells . let @xmath316 be a compact @xmath7-surjective packed subcomplex of a cover @xmath8 . if @xmath316 is not @xmath7-injective , then there is a minimal area disc diagram @xmath619 whose boundary cycle is an essential immersed path in @xmath316 . as in the proof of theorem [ thm : fan - diagramct ] , a minimal area disc diagram for an essential immersed path in @xmath4 yields a sequence of weakly perimeter reducing spread - out fans that can be attached . we claim that in the appropriate sense @xmath620 is missing along the @xmath47-cells in @xmath621 that map to @xmath47-cells of @xmath326 in @xmath202 . indeed , if some @xmath0-cell of @xmath620 was already contained in @xmath326 then a corresponding @xmath0-cell @xmath50 of @xmath574 is contained in @xmath326 . let @xmath87 be the concatenation @xmath622 where @xmath623 is the part of @xmath87 that is the subpath of the outer path @xmath73 of @xmath574 . now @xmath386 is homotopic in @xmath624 and @xmath326 to a path @xmath625 which is identical to @xmath69 except that @xmath626 is substituted for @xmath623 . since @xmath625 does nt go around @xmath627 , we see that @xmath628 and therefore after removing spurs from a disc diagram for @xmath625 ( and identifying some @xmath47-cells on the boundary ) , we obtain an immersed path homotopic to @xmath386 in @xmath326 whose area is strictly less than the area of @xmath386 which is impossible . we let @xmath325 be the union of @xmath326 with ( the image of ) @xmath620 be a new compact subcomplex in @xmath8 . now the outer path @xmath73 of @xmath574 extends to a path @xmath629 , and the argument breaks down according to whether @xmath621 is contained in @xmath326 . if @xmath630 then our hypothesis that @xmath631 implies that @xmath356 . if @xmath621 is not a path in @xmath326 , then @xmath69 is not null - homotopic in @xmath325 since @xmath632 and @xmath633 collapses onto the union of @xmath326 and some nontrivial arcs . now our hypothesis that @xmath574 is weakly perimeter reducing implies that @xmath634 , and @xmath325 contains the essential immersed path @xmath385 with @xmath635 , where @xmath385 is defined as follows : first remove the interiors of @xmath574 and @xmath73 from @xmath624 to obtain a diagram @xmath636 , and then fold @xmath637 until the boundary is immersed . note that we can not obtain a sphere in this way , because otherwise @xmath621 is the same as @xmath638 and so we could have used @xmath574 instead of @xmath624 to begin with , contradicting that @xmath624 is minimal area . this process can only be repeated finitely many times without the perimeter strictly decreasing and hence @xmath8 satisfies the perimeter reduction hypothesis and so @xmath14 is coherent . in this section we review the interconnections between quasi - isometries , quasiconvexity , and word - hyperbolicity . since these results are well - known , we simply state the definitions and lemmas we will need and refer the interested reader to @xcite , @xcite , and @xcite for more detailed accounts . [ def : geo ] let @xmath639 and @xmath640 be metric spaces . a map @xmath641 which preserves distances is called an _ isometric embedding _ of @xmath642 into @xmath6 , and an isometric embedding of an interval @xmath643 $ ] of the real line is called a _ geodesic _ from @xmath644 to @xmath645 . if any two points in @xmath6 can be connected by a geodesic , then @xmath6 is a _ geodesic metric space_. a fundamental example of a geodesic metric space is a connected graph with the path metric . note that by a ` graph ' we mean a @xmath47-dimensional cw - complex , so that loops and multiple edges are allowed . [ def : graphs ] the _ path metric _ on a connected graph @xmath646 makes each @xmath47-cell of @xmath646 locally isometric to the unit interval , and then defines the distance between two arbitrary points of @xmath646 to be the length of the shortest path between them . it is easy to see that such a minimal path always exists , and that it will be a geodesic . thus connected graphs are geodesic metric spaces using the path metric . [ def : cayley ] let @xmath6 be a connected @xmath0-complex , and let @xmath647 be its universal cover . since the @xmath47-skeleton @xmath648 is a connected graph it is a geodesic metric space with the path metric . if @xmath6 has a unique @xmath21-cell , then @xmath6 is the standard @xmath0-complex of some group presentation @xmath649 , and the graph @xmath648 is the _ cayley graph of the presentation_. alternatively , the cayley graph , often denoted @xmath22 , can be defined as follows . begin with a @xmath21-cell set corresponding to the elements of @xmath19 and an edge set labeled by the elements of @xmath650 . then attach the edges to the @xmath21-cells so that the edge labeled @xmath651 begins at the @xmath21-cell @xmath652 and ends at the @xmath21-cell @xmath653 . we endow @xmath22 with the path metric . since @xmath19 can be identified with the @xmath21-skeleton of @xmath22 , we can metrize @xmath19 by giving it the subspace metric . since this metric on @xmath19 depends on the generating set @xmath20 , we will denote the resulting metric space by @xmath654 . although distinct generating sets for @xmath19 will produce distinct metrics using this procedure , all of the metrics on a finitely generated group will be roughly equivalent . we will now make this precise . [ def : qgeo ] let @xmath639 and @xmath640 be metric spaces and let @xmath641 be a map between them . if there exist constants @xmath655 and @xmath656 such that for all @xmath657 , @xmath658 then @xmath58 is a _ @xmath659-quasi - isometric embedding _ of @xmath642 into @xmath6 . the special case of a @xmath659-quasi - isometric embedding of an interval of the real line into @xmath6 is a _ @xmath659-quasigeodesic_. if every point in @xmath6 is within a uniformly bounded distance of a point in the image of @xmath58 , then @xmath58 is a _ @xmath659-quasi - isometry _ between @xmath642 and @xmath6 . a map will be called a _ quasi - isometry _ if it is a @xmath659-quasi - isometry for some choice of @xmath660 and @xmath17 , and the spaces involved will be said to be quasi - isometric . the notion of quasi - isometry is an equivalence relation on spaces in the following sense . if there is a quasi - isometry from @xmath6 to @xmath4 then there also exists a quasi - isometry from @xmath4 to @xmath6 , and if @xmath661 and @xmath662 are quasi - isometries , then the composition @xmath102 is also a quasi - isometry . if a finitely generated group @xmath19 acts in a reasonably nice way on a reasonably nice space , then the group , using the metric derived from its cayley graph , will be quasi - isometric to the space it acts on . the following theorem contains the precise statement of this fact . [ lem : nice - action ] let @xmath6 be a locally compact , connected , geodesic metric space . let @xmath19 be a finitely generated group which acts on @xmath6 properly discontinuously and cocompactly by isometries . then for any point @xmath663 , and for any finite set of generators @xmath664 , the map @xmath665 defined by @xmath666 is a quasi - isometry , where we give @xmath19 its cayley graph metric relative to the generating set @xmath20 . as a corollary , we see that changing generating sets induces a quasi - isometry . [ cor : qi ] if @xmath20 and @xmath96 are finite generating sets for a group @xmath19 , then the metric spaces @xmath654 , @xmath667 , @xmath22 , and @xmath668 are all quasi - isometric . furthermore , the quasi - isometry @xmath669 is induced by the identity map @xmath670 . a second fundamental notion is that of a quasiconvex subspace of a metric space . [ def : qcon ] a subspace @xmath4 of a geodesic metric space @xmath6 is _ @xmath660-quasiconvex _ if there is a @xmath660-neighborhood of @xmath4 which contains all of the geodesics of @xmath6 that begin and end in @xmath4 . a subspace is @xmath671 provided that it is @xmath660-quasiconvex for some @xmath660 . the notion of quasiconvexity can be extended to groups and subgroups via cayley graphs . specifically , a subgroup @xmath18 of a group @xmath19 generated by @xmath20 is _ quasiconvex _ if the @xmath21-cells corresponding to @xmath18 form a quasiconvex subspace of @xmath22 . the group @xmath19 generated by @xmath20 is _ locally quasiconvex _ if every finitely generated subgroup is quasiconvex . we record the following two properties of quasiconvex subgroups . see @xcite and the references therein for details . [ lem : short ] if @xmath18 is a quasiconvex subgroup of a group @xmath19 generated by finite set @xmath20 , then @xmath18 itself is finitely generated . [ lem : qc->qie ] let @xmath19 be a group with finite generating set @xmath20 and let @xmath18 be a subgroup of @xmath19 with finite generating set @xmath96 . if @xmath18 is a quasiconvex subspace of @xmath22 , then @xmath672 is quasi - isometrically embedded in @xmath22 . although the various metrics which have been defined for a group @xmath19 are all equivalent up to quasi - isometry ( corollary [ cor : qi ] ) , the generating set @xmath20 does need to be specified in definition [ def : qcon ] . this is because the notion of quasiconvexity is not well - behaved under quasi - isometries . in particular , the group @xmath673 shows that the converse of lemma [ lem : qc->qie ] is false . the dependence of quasiconvexity on generating sets and the distinction between quasiconvex subgroups and quasi - isometrically embedded subgroups disappears once we restrict our attention to word - hyperbolic groups . [ def : hyperbolic ] let @xmath80 , @xmath81 , and @xmath135 be points in a geodesic metric space @xmath6 and let @xmath474 be a triangle of geodesics connecting @xmath80 to @xmath81 , @xmath81 to @xmath135 and @xmath80 to @xmath135 . this geodesic triangle is _ @xmath674-thin _ if each of the sides is contained in a @xmath674-neighborhood of the union of the other two . if there is a uniform @xmath674 such that every geodesic triangle in @xmath6 is @xmath674-thin , then @xmath6 is a _ @xmath674-hyperbolic space_. a group @xmath19 generated by a finite set @xmath20 is _ word - hyperbolic _ if its cayley graph @xmath22 is @xmath674-hyperbolic . one of the key properties of @xmath674-hyperbolic spaces is that geodesics and quasigeodesics stay uniformly close in the following sense : [ lem : close ] let @xmath80 and @xmath81 be points in the @xmath674-hyperbolic metric space @xmath6 . then there are integers @xmath675 and @xmath676 such that if @xmath520 is a @xmath677-quasigeodesic between the points @xmath678 and @xmath679 is a geodesic @xmath680 $ ] , then @xmath679 is contained in an @xmath681-neighborhood of @xmath520 and @xmath520 is contained in an @xmath682-neighborhood of @xmath679 . it is easy to deduce from lemma [ lem : close ] that the property of being @xmath674-hyperbolic for some @xmath674 is preserved by quasi - isometries between geodesic metric spaces , even though the specific value of @xmath674 may have to be changed . combined with corollary [ cor : qi ] , this shows that the property of a group being word - hyperbolic is independent of the choice of a finite generating set . [ cor : preserves quasi ] if @xmath6 and @xmath642 are geodesic metric spaces , @xmath6 is @xmath674-hyperbolic , and @xmath683 is a quasi - isometry , then a subspace @xmath4 is quasiconvex in @xmath6 if and only if @xmath684 is quasiconvex in @xmath642 . as a consequence , the quasiconvexity of a subgroup in a word - hyperbolic group does not depend on the generating set . [ cor : independent ] let @xmath18 be a subgroup of the word - hyperbolic group @xmath19 and let @xmath20 and @xmath96 be finite generating sets for @xmath19 . the subgroup @xmath18 will be quasiconvex in @xmath22 if and only if @xmath18 is quasiconvex in @xmath668 . in particular , when @xmath19 is word - hyperbolic , the quasiconvexity of a subgroup is independent of the choice of finite generating set for @xmath19 . thus , for word - hyperbolic groups there is the following partial converse to lemma [ lem : qc->qie ] . [ lem : qie->qc ] let @xmath19 be a word - hyperbolic group with finite generating set @xmath20 and let @xmath18 be a subgroup of @xmath19 with finite generating set @xmath96 . if @xmath672 is quasi - isometrically embedded in @xmath654 , then @xmath18 is a quasiconvex subgroup of @xmath19 . in this section we prove our main technical results about local quasiconvexity . since the reader has already seen arguments utilizing perimeter - reducing fans in section [ sec : fanct ] , we will treat only the general fan case . the reader not yet completely comfortable with the language of fans should keep in mind the special case of a fan consisting of a single @xmath0-cell @xmath50 with outer path @xmath73 in @xmath87 . [ def : qrh ] let @xmath6 be a weighted @xmath0-complex , let @xmath590 be a collection of fans in @xmath6 , and let @xmath660 and @xmath17 be constants . a path @xmath212 can be _ @xmath659-straightened _ if there exists a sequence of paths @xmath685 such that for each @xmath40 , @xmath385 is obtained from @xmath386 by either removing a backtrack , or by pushing across a fan in @xmath590 . in addition , the final path @xmath384 must satisfy the following condition : consider the lift of @xmath384 to @xmath122 and let @xmath293 denote the length of a geodesic in @xmath686 with the same endpoints . there must exist a path @xmath687 with the same endpoints as @xmath688 such that @xmath220 lies in a @xmath660-neighborhood of @xmath384 and such that @xmath689 if every fan in @xmath590 is spread - out and perimeter - reducing , and if for some fixed choice of @xmath660 and @xmath17 , every path @xmath212 can be @xmath659-straightened , we say that @xmath6 satisfies the _ straightening hypothesis_. the following is our main technical result about the straightening hypothesis . [ thm : lqc ] let @xmath6 be a compact weighted @xmath0-complex . if @xmath6 satisfies the straightening hypothesis , then every finitely generated subgroup of @xmath14 embeds by a quasi - isometry . furthermore , if @xmath14 is word - hyperbolic then it is locally quasiconvex . let @xmath690 , let @xmath18 be a subgroup of @xmath19 which is generated by a finite set @xmath96 of elements , and let @xmath8 be the based covering space of @xmath6 corresponding to the inclusion @xmath691 . next , let @xmath13 be a wedge of finitely many circles , one for each generator in @xmath96 , and let @xmath692 be a map which sends each circle to a based path in @xmath437 representing its corresponding generator . this map lifts to a map @xmath693 , and we let @xmath694 denote the image of @xmath13 in @xmath8 . since the weights on the sides of the @xmath0-cells are nonnegative integers , the perimeter of any compact subcomplex of @xmath8 is finite and nonnegative . in particular , there exists some compact connected subcomplex @xmath695 which contains @xmath694 and which does not admit any fan perimeter reductions . for instance , we can choose @xmath4 to be of minimal perimeter among all compact connected subcomplexes of @xmath8 containing @xmath694 . if @xmath379 were to admit a fan perimeter reduction , then lemma [ lem : fan - attachments ] and lemma [ lem : surj ] would allow us to create a slightly larger subcomplex which had a strictly smaller perimeter , contradicting the way @xmath4 was chosen . let @xmath66 be a @xmath660-neighborhood of @xmath4 in @xmath8 , let @xmath696 be the based universal cover of @xmath6 , let @xmath210 denote the based component of the preimage of @xmath4 in @xmath647 , and let @xmath209 denote the based connected component of the preimage of @xmath66 in @xmath647 . since @xmath697 , we see that @xmath66 contains a set of paths which generate @xmath698 and thus the action of @xmath691 on @xmath647 stabilizes @xmath209 . in particular , the preimages of the basepoint of @xmath8 in @xmath647 are contained in @xmath209 and these @xmath21-cells are in one - to - one correspondence with the elements of @xmath18 . using this correspondence , we will consider @xmath18 as a subspace of @xmath647 . let @xmath699 be the metric on @xmath18 defined by the @xmath47-skeleton of @xmath700 . specifically , define @xmath701 to be the length of the shortest path in @xmath209 between the appropriate @xmath21-cells of @xmath18 . since @xmath4 does not admit any fan perimeter reductions and @xmath66 is the @xmath660-neighborhood of @xmath4 , the straightening hypothesis allows us to conclude that every pair of points in @xmath702 is connected by a path in @xmath209 which satisfies equation ( [ eq : length - condition ] ) . in particular , given a path @xmath69 in @xmath210 connecting a pair of points in @xmath18 , we can follow the sequence of alterations to obtain paths @xmath703 without leaving the subcomplex @xmath210 , and since the path @xmath220 lies in a @xmath660-neighborhood of @xmath384 , the path @xmath220 does not leave the subcomplex @xmath704 . finally , since this is true for all pairs of points in @xmath702 , this shows that the map @xmath705 is a @xmath659-quasi - isometric embedding . since @xmath6 is compact , @xmath19 has some finite generating set @xmath20 . more specifically , if we select a maximal spanning tree for @xmath437 , then a generator corresponding to each @xmath47-cell not in the spanning tree is sufficient . consider the following diagram of maps between metric spaces where the metric spaces @xmath654 and @xmath22 are the metric on the group and the metric on its cayley graph . @xmath706 we have shown that the bottom map is a quasi - isometric embedding . since @xmath19 acts properly discontinuously and cocompactly on @xmath707 , by lemma [ lem : nice - action ] the map @xmath708 is a quasi - isometry . next , as remarked above , the action of @xmath691 on @xmath647 stabilizes @xmath209 . the action of @xmath18 on @xmath704 is clearly properly discontinuous since it is a restriction of @xmath19 on @xmath647 , and it is cocompact since the quotient of @xmath709 by @xmath18 is the compact space @xmath66 . thus , by lemma [ lem : nice - action ] the map @xmath710 is also a quasi - isometry . combining these three maps we see that the map @xmath711 is a quasi - isometric embedding . since by corollary [ cor : qi ] , @xmath712 is also a quasi - isometry , the map from @xmath672 to @xmath22 is a quasi - isometric embedding as well . finally , if @xmath19 is word - hyperbolic then it follows from lemma [ lem : qie->qc ] that @xmath18 is quasiconvex . we are unable to answer the following problem about the relationship between the straightening hypothesis and word - hyperbolicity . however , one can show the answer is affirmative if one adds to definition [ def : qrh ] the requirement that @xmath384 lie in a @xmath660-neighborhood of @xmath220 . [ prob : qrh->hyp ] suppose the compact weighted @xmath0-complex @xmath6 satisfies the straightening hypothesis with respect to some finite collection @xmath590 of fans in @xmath6 . is @xmath14 word - hyperbolic ? two conditions which immediately imply the straightening hypothesis are a decrease in length and a decrease in area . more explicitly , if for every immersed path @xmath68 which does not lift to a @xmath659-quasigeodesic , the path @xmath69 can be pushed across a perimeter - reducing fan to obtain a new path of strictly smaller length , then @xmath6 will satisfy the straightening hypothesis . indeed the sequence of reductions can not continue indefinitely because the length decreases each time , and thus they terminate at a path @xmath713 which lifts to a quasigeodesic . the most important condition that implies the straightening hypothesis will involve the following notion : [ def : j - thin ] a disc diagram @xmath387 with boundary cycle @xmath548 is called _ @xmath714-thin _ for some @xmath715 , if every @xmath21-cell in @xmath69 is contained in a @xmath714-neighborhood of @xmath73 and vice - versa . [ thm : fan - diagramlqc ] let @xmath590 be a finite collection of perimeter - reducing spread - out fans in the compact weighted @xmath0-complex @xmath6 , and let @xmath716 . suppose that for every minimal area disc diagram @xmath717 with boundary cycle @xmath548 , either @xmath387 is @xmath714-thin or @xmath387 contains a spur or fan in @xmath590 whose outer path is a subpath of either @xmath69 or @xmath73 . then @xmath14 is a locally quasiconvex word - hyperbolic group . we will first give an argument that shows that @xmath6 satisfies the straightening hypothesis . we will then apply a special case of this argument to see that @xmath14 is word - hyperbolic . the result will then follow from theorem [ thm : lqc ] . let @xmath213 be a path , let @xmath718 be a lift of this path to the universal cover @xmath647 , and let @xmath719 be a geodesic in @xmath647 with the same endpoints . since the path @xmath548 is a closed null - homotopic path in @xmath647 , its projection to @xmath6 has the same properties . let @xmath389 be a minimal area disc diagram with boundary cycle @xmath548 . let @xmath387 be oriented so that the path @xmath69 proceeds from left to right across the top of the diagram and the path @xmath73 proceeds from left to right across the bottom . using the diagram @xmath387 , we will now construct an explicit sequence of paths @xmath386 demonstrating the straightening hypothesis . along the way we will need to define a sequence of paths @xmath581 , a sequence of diagrams @xmath624 , and a sequence of diagrams @xmath720 as well . the idea will be to systematically remove portions of @xmath387 from the top and bottom . at each stage of this process the paths along the top and bottom will be @xmath386 and @xmath581 , the diagram between @xmath386 and @xmath581 will be @xmath624 , and the diagram bounded by @xmath721 will be @xmath720 . at the end we will reach a diagram @xmath722 with boundary paths @xmath723 and @xmath724 such that @xmath725 is the boundary cycle of @xmath722 , and @xmath722 is @xmath714-thin . at each stage @xmath624 and @xmath720 will be subdiagrams of @xmath387 . the reader is referred to figure [ fig : lqcdiagram ] for an illustration of the diagram @xmath387 as well as some of the relevant paths and subdiagrams appearing in the final situation . of @xmath387 illustrated above is @xmath47-thin . [ fig : lqcdiagram ] ] to begin the process , let @xmath726 , @xmath727 , @xmath728 , and let @xmath729 be the diagram without @xmath0-cells consisting of the path @xmath73 . this is a diagram since @xmath73 is a geodesic in @xmath647 and thus simple in @xmath387 . since definition [ def : qrh ] permits the removal of backtracks , we may assume that our initial path @xmath391 is immersed . now for each @xmath40 , since both @xmath730 and @xmath731 are immersed paths , by hypothesis the diagram @xmath624 is either @xmath714-thin , or @xmath624 contains a fan @xmath732 whose boundary path is a subpath of either @xmath386 or @xmath581 . removing the interior of the fan and the interior of its boundary path from @xmath624 defines a new connected and simply - connected diagram @xmath733 where the new boundary paths across the top and bottom of @xmath733 are @xmath385 and @xmath734 . notice that by hypothesis , the path @xmath385 is either identical to @xmath386 or it is obtained from @xmath386 by a perimeter - reducing push across a fan , and likewise , @xmath734 is either identical to @xmath581 or it is obtained from @xmath581 by a perimeter - reducing push across a fan . in either case , the fact that the original path is an immersion implies that the new path is an immersion as well . since the total number of cells in the diagrams @xmath624 is decreasing , this process must eventually terminate at a diagram @xmath722 which is @xmath714-thin . it only remains to show that the final path @xmath384 lies close to a path @xmath220 which satisfies the length condition mentioned in definition [ def : qrh ] . let @xmath735 be a bound on the perimeter of a @xmath47-cell in @xmath6 ( which is compact ) , let @xmath561 be a bound on the lengths of inner paths of fans in @xmath590 , and let @xmath660 be the larger of @xmath714 and @xmath736 . let @xmath737 and note that since @xmath722 is @xmath714-thin , @xmath220 lies in a @xmath660-neighborhood of @xmath384 . we will now show that @xmath738 . let @xmath739 and @xmath740 be the packed versions of @xmath741 and @xmath742 . now since @xmath734 is obtained from @xmath581 by a perimeter - reducing push across a fan , we have @xmath743 , and consequently , @xmath744 . finally , since @xmath745 for each @xmath40 , we have @xmath746 . we will prove that @xmath14 is word - hyperbolic by showing that @xmath14 has a linear isoperimetric function @xcite . we will apply the above argument except that we will exchange the roles of @xmath69 and @xmath73 , so that @xmath73 is an arbitrary immersed null - homotopic path and @xmath69 is the trivial path . first observe that since no outer path of a fan can be a subpath of @xmath69 , we see that @xmath747 . we will now use the fact that @xmath748 to estimate @xmath749 . let @xmath750 be the maximal number of @xmath0-cells occurring in a fan in @xmath590 . then @xmath751 was obtained from @xmath752 by adding at most @xmath753 fans , and so @xmath754 . since the diagram @xmath722 is @xmath714-thin , the path @xmath755 is contained in a @xmath714-neighborhood of the path @xmath756 . letting @xmath13 denote the ball of radius @xmath714 in @xmath647 , lemma [ lem : cell - bound ] ( established below ) implies that @xmath757 where @xmath682 is a constant that depends only on @xmath6 and @xmath714 . but @xmath758 so @xmath759 , and the isoperimetric function is linear as claimed . [ lem : cell - bound ] if @xmath647 is a simply - connected @xmath0-complex and @xmath13 is a compact , connected subspace of @xmath647 , then there is a constant @xmath682 , depending only on @xmath13 , such that for every closed path @xmath760 , we have @xmath761 . let @xmath762 be the finitely many simple closed nontrivial paths in @xmath13 , and for @xmath763 let @xmath764 where @xmath765 is a minimal area disc diagram with boundary cycle @xmath766 . such a @xmath624 exists for each @xmath766 since @xmath6 is simply - connected . let @xmath682 be the maximum value of @xmath767 . intuitively , @xmath682 measures the maximum number of @xmath0-cells needed per @xmath47-cell in a simple closed path . the necessary inequality is now easy to establish by breaking any closed path @xmath760 into simple closed paths and backtracks , creating disc diagrams for each individually , and then reassembling them into a disc diagram @xmath387 for the original path . since the maximum number of @xmath0-cells needed per @xmath47-cell is bounded by @xmath682 for each individual portion , @xmath768 as claimed . we close this section with a problem analogous to problem [ prob : coherence ] : [ question : lqc ] let @xmath6 be a compact weighted @xmath0-complex such that @xmath14 is word - hyperbolic . let @xmath660 and @xmath17 be fixed constants , and suppose that for every path @xmath213 , there exists a @xmath659-quasigeodesic @xmath73 and a disc diagram @xmath474 with boundary cycle @xmath548 , such that @xmath769 . is @xmath14 locally quasiconvex ? in this section we apply the local quasiconvexity result to various small cancellation groups . [ thm : lqc - sc-1 ] let @xmath6 be a weighted @xmath0-complex which satisfies @xmath549-@xmath550 @xmath551-@xmath552 $ ] . suppose @xmath770 for each @xmath0-cell @xmath113 and path @xmath485 which is the concatenation at most three @xmath544two@xmath546 $ ] consecutive pieces in the boundary of @xmath50 . then @xmath14 is coherent . this follows immediately from corollary [ cor : fan - classification ] and theorem [ thm : fan - diagramlqc ] . [ surfacegps ] let @xmath6 be the standard @xmath0-complex of the presentation @xmath771 . then @xmath6 is the usual cell structure for the nonorientable surface of genus @xmath652 . clearly for any @xmath772 , @xmath6 satisfies @xmath773-@xmath537 , and the pieces are of length @xmath47 . using the unit perimeter we see that the weight of any piece ( @xmath47-cell ) is @xmath0 and the weight of the @xmath0-cell is @xmath774 . thus by theorem [ thm : coh - sc-1 ] , @xmath14 is coherent for @xmath775 and by theorem [ thm : lqc - sc-1 ] , @xmath14 is locally quasiconvex for @xmath776 . a similar result holds if we let @xmath6 be the standard @xmath0-complex of the presentation @xmath777[a_2,b_2 ] \ldots [ a_g , b_g ] \rangle\ ] ] so that @xmath6 is an orientable surface of genus @xmath652 . for @xmath778 the @xmath0-complex @xmath6 satisfies @xmath779-@xmath537 , the pieces are of length @xmath47 , the weight of each piece is @xmath0 , and the weight of the @xmath0-cell is @xmath780 . thus by theorem [ thm : coh - sc-1 ] , @xmath14 is coherent for @xmath778 and by theorem [ thm : lqc - sc-1 ] , @xmath14 is locally quasiconvex for @xmath781 . the fact that these methods can be used to prove the coherence and local quasiconvexity of surface groups is to be expected since the boundary of a @xmath0-manifold was one of the original motivations for the notion of perimeter introduced in section [ sec : perimeter ] . here is a more novel application of theorem [ thm : coh - sc-1 ] . [ thm : few - occurrences ] let @xmath41 be a presentation that satisfies @xmath42 . if each @xmath43 occurs at most @xmath44 times among the @xmath45 , then @xmath19 is coherent and locally quasiconvex . in case @xmath782 this is obvious , because any generator appears at most @xmath783 times , and consequently at most once , which implies that the group is free . on the other hand , when @xmath784 , the complex @xmath6 satisfies @xmath549 . use the unit weighting , and let @xmath159 be a subpath of the @xmath0-cell @xmath45 consisting of three consecutive pieces . the small cancellation assumption implies that @xmath785 . on the other hand , the bound on the number of occurrences of each generator shows that @xmath786 for each @xmath47-cell @xmath55 . thus @xmath787 . by theorem [ thm : coh - sc-1 ] the group is coherent and by theorem [ thm : lqc - sc-1 ] it is locally quasiconvex . in section [ sec:1relator ] we showed that certain one - relator groups with torsion are coherent . since it was shown in @xcite that @xmath788 satisfies the @xmath789 small cancellation condition , we can apply theorem [ thm : lqc - sc-1 ] to obtain a local - quasiconvexity result as well . [ thm : lqc - one - rel ] let @xmath27 be a one - relator group with @xmath28 . then @xmath19 is locally quasiconvex . let @xmath6 be the @xmath0-complex of this presentation with the unit weighting . since the weight of the unique @xmath0-cell is @xmath790 , the weight of its packet is @xmath791 . on the other hand , by the spelling theorem ( theorem [ thm : spelling ] ) , the length of a piece is less than @xmath232 and we can assume that each generator appears in @xmath23 fewer than @xmath232 times , for otherwise the group is virtually free and the theorem is obvious . consequently , the perimeter of each @xmath47-cell is at most @xmath792 and since @xmath793 , this is bounded by @xmath794 . so the perimeter of a piece is strictly less than @xmath795 . thus for @xmath28 , @xmath19 is locally quasiconvex by theorem [ thm : lqc - sc-1 ] . a more detailed examination of the local quasiconvexity of one - relator groups with torsion has been carried out by hruska and wise in @xcite . as an application of theorem [ thm : lqc ] they are able to prove the following : let @xmath796 be a one - relator group with @xmath797 . then @xmath19 is locally quasiconvex . our next application is to finitely presented small cancellation groups with torsion . as in the one - relator case , small cancellation groups with sufficient torsion will always be coherent and locally quasiconvex . in the proof we will need the following classical lemma about words in the free group ( see @xcite ) . [ lem : x+y ] let @xmath6 and @xmath4 be cyclically reduced words in the free group which are not proper powers . if @xmath6 and @xmath4 are not cyclic conjugates , and a word @xmath477 is both a subword of a power of @xmath6 and a subword of a power of @xmath4 , then @xmath798 . [ thm : powers ] let @xmath29 be a finite presentation , where each @xmath30 is a cyclically reduced word which is not a proper power . if @xmath30 is not freely conjugate to @xmath31 for @xmath32 , then there exists a number @xmath33 such that for all choices of integers @xmath34 the group @xmath35 is coherent . specifically , the number @xmath799 has this property , where @xmath37 and @xmath38 denote longest and shortest words among the @xmath30 , respectively . moreover , if @xmath39 for all @xmath40 , then @xmath19 is locally quasiconvex . we will assume that @xmath800 . essentially , the only case that this assumption eliminates is the presentation @xmath801 and the presentation @xmath802 , and the theorem is trivial in these cases . let @xmath33 be the number satisfying equation ( [ eq : powers ] ) and choose @xmath34 for all @xmath40 . we will first show that @xmath803 satisfies the @xmath549 condition . by lemma [ lem : x+y ] the length of the longest piece between the @xmath0-cell labeled @xmath804 and the @xmath0-cell labeled @xmath805 is bounded above by @xmath806 . notice that @xmath807 is thus a uniform bound on the size of a piece which is independent of the size of the chosen @xmath808 . since the length of the @xmath40-th @xmath0-cell is @xmath809 , and since by assumption , latexmath:[\[n_i \cdot |w_{\text{min}}| \geq n \cdot |w_{\text{min}}| = 6 \cdot |w_{\text{max}}| \sum |w_i| \geq 6 \cdot 2 \cdot we can conclude that the @xmath549 condition is satisfied . next we will choose a weighting on the sides of the @xmath0-cells of the standard @xmath0-complex for @xmath811 , and verify that the weight criterion of theorem [ thm : coh - sc-1 ] is satisfied . let @xmath812 and assign a weight of @xmath813 to each of the sides of the @xmath0-cell @xmath814 corresponding to the relator @xmath804 . if we let @xmath159 denote a path in @xmath582 consisting of at most three consecutive pieces , then we must show that @xmath815 . this follows from the following string of inequalities : @xmath816 the first inequality is true since @xmath817 , which is clearly greater than or equal to @xmath818 . the middle inequality uses the restrictions assumed on @xmath808 and @xmath33 . finally , the last inequality is a combination of two observations : ( 1 ) the perimeter of a single @xmath47-cell will be at most the sum of the weights of all of the sides in the complex , so that @xmath819 . together these show the final inequality . the weight criterion of theorem [ thm : coh - sc-1 ] is thus satisfied and the group @xmath820 is coherent . similarly , if the inequalities are strict , then by theorem [ thm : lqc - sc-1 ] , it is locally quasiconvex . although there is some overlap between the groups studied in this article and those studied by feighn and handel in @xcite , the methods and the results are distinct . one indication of this is that all of the groups in @xcite are indicable ( i.e. admit a homomorphism onto @xmath2 ) , whereas theorem [ thm : powers ] can be used to construct coherent groups which are perfect . [ cor : perfect ] there exist perfect groups which satisfy the perimeter reduction hypothesis and are thus coherent . the following example illustrates the idea : consider the presentation @xmath821 if @xmath33 is chosen to satisfy equation ( [ eq : powers ] ) , then the following group is coherent by theorem [ thm : powers ] . @xmath822 but the following presentation for the abelianization of @xmath19 shows that it is trivial : @xmath823 , a^n , a^{n+1 } , b^n , b^{n+1}\rangle\ ] ] in this section , we use the theorems about coherence and local quasiconvexity in small cancellation groups to show that a large family of @xmath1-manifold groups are coherent and locally quasiconvex . we begin with a theorem about branched covers of @xmath0-complexes . [ thm : branched 2-complex ] let @xmath6 be a compact @xmath0-complex , and suppose that no @xmath0-cell of @xmath6 is attached by a proper power . then there exists a constant @xmath293 depending only on @xmath6 such that for every branched cover @xmath824 where ( 1 ) the branching is over the barycenters of @xmath0-cells of @xmath6 , and ( 2 ) all of the branching degrees are at least @xmath293 , the fundamental group @xmath16 is coherent . similarly , there is another constant @xmath293 such that the compact branched covers satisfying these conditions have a locally quasiconvex fundamental group . the proof is similar to the proof of theorem [ thm : powers ] , which can actually be deduced from it . first notice that it is sufficient to consider the case where no two @xmath0-cells have the same attaching maps . it follows from lemma [ lem : x+y ] that there is a bound depending only on @xmath6 on the length of a piece in the boundary of a @xmath0-cell of @xmath8 . thus the lengths of the pieces remain bounded as the branching degrees increase , whereas the length of the boundary of each @xmath0-cell grows linearly with @xmath293 . it follows that for large @xmath293 the presentation satisfies small cancellation conditions . using the unit perimeter , the maximum sum of the weight of three consecutive pieces remains constant , whereas the weight of each @xmath0-cell grows linearly with @xmath293 . thus , for sufficiently large @xmath293 the weight criteria for coherence ( theorem [ thm : coh - sc-1 ] ) and for local quasiconvexity ( theorem [ thm : lqc - sc-1 ] ) will be satisfied . [ rem : numerous branched ] observe that finite branched covers with high branching degrees correspond to certain finite index subgroups of the fundamental group of the space we obtain when we remove the barycenter of each @xmath0-cell . since the fundamental group of this space is free , and thus residually finite ( @xcite ) , these types of covers are numerous . [ thm : branched manifold ] let @xmath682 be a compact @xmath1-manifold with a combinatorial cell structure . there exists @xmath293 depending on @xmath682 such that the following holds : let @xmath825 be a branched cover with at least @xmath293 fold branching along each @xmath47-cell of @xmath682 . then @xmath826 is coherent or even locally quasiconvex . let @xmath827 denote the underlying manifold of @xmath682 equipped with the cell structure dual to @xmath682 . the branched covers @xmath828 branched along @xmath47-cells of @xmath682 correspond to branched covers @xmath829 branched along centers of @xmath0-cells of @xmath830 . note that @xmath831 is a subspace of @xmath96 . furthermore , the branching degree at each @xmath47-cell of @xmath682 is the same as the branching degree at the center of the corresponding @xmath0-cell of @xmath830 . the obvious strong deformation retraction of @xmath832 onto @xmath830 induces a strong deformation retraction of @xmath96 onto @xmath831 , so that @xmath833 . it is therefore sufficient to prove the analogous result for branched covers @xmath834 where the branching occurs over the centers of @xmath0-cells of @xmath830 , and this is exactly what was proved in theorem [ thm : branched 2-complex ] . we illustrate theorem [ thm : branched manifold ] with the following example . [ exmp : branched cube ] let @xmath682 denote the usual cell division for the @xmath1-torus @xmath835 . we will show that any branched cover of @xmath836 along the @xmath47-cells of @xmath682 has a coherent fundamental group provided that the degree @xmath293 of branching is @xmath515 . first observe that since @xmath827 is obviously isomorphic to @xmath682 , it is easy to see that @xmath830 contains exactly three @xmath0-cells , each of which is a square . each piece of @xmath830 has length @xmath47 and perimeter @xmath378 . if each branching degree is @xmath837 then the presentation satisfies @xmath838 . since the perimeter of three consecutive pieces is @xmath839 and the weight of each @xmath0-cell at least @xmath840 , the fundamental group will be coherent when @xmath841 ( by theorem [ thm : coh - sc-1 ] ) and locally quasiconvex when @xmath842 ( by theorem [ thm : lqc - sc-1 ] ) . as mentioned in the introduction , scott and shalen proved that all @xmath1-manifold groups are coherent , so the coherence assertion in theorem [ thm : branched manifold ] is certainly not new . nevertheless , it is interesting to be able to recover this special case from the different point of view of this paper . the local quasiconvexity assertion is a bit trickier to obtain using prior results . however , it seems likely that the branched covers of theorem [ thm : branched manifold ] can be constructed from hyperbolic @xmath1-manifolds with non - empty boundary and no cusps , by gluing along annuli . thus the local quasiconvexity appears to follow from the following theorem of thurston s @xcite which we quote from ( * ? ? ? * proposition 7.1 ) , together with the theorem due to swarup @xcite that in the case that @xmath33 has no cusps , geometrically finite subgroups are quasiconvex . [ thm : atoroidal ] let @xmath33 be a geometrically finite hyperbolic manifold such that @xmath843 is nonempty . then every covering space @xmath844 of @xmath33 with a finitely generated fundamental group is also geometrically finite . we believe that the branched covers with branching degree @xmath845 of the @xmath1-torus in example [ exmp : branched cube ] are atoroidal hyperbolic @xmath1-manifolds with boundary , and hence the local quasiconvexity does follow from theorem [ thm : atoroidal ] . there are a number of properties of groups which are closely related to coherence and local quasiconvexity . in this final section we examine briefly how our techniques can be used to produce results about three of these related topics : howson s property , finitely generated intersections with magnus subgroups , and the generalized word problem . a group is said to satisfy the _ finitely generated intersection property _ ( or f.g.i.p . ) if the intersection of any two finitely generated subgroups is also finitely generated . in 1954 howson proved that free groups have the f.g.i.p . , and as a result this property is sometimes referred to as howson s property . as was shown in @xcite , every quasiconvex subgroup is finitely generated and the intersection of any two quasiconvex subgroups is again quasiconvex . combining these two facts , one sees that every locally quasiconvex group satisfies the f.g.i.p . in particular , all of the groups we have shown to be locally quasiconvex , also have howson s property . in this subsection we show how the algorithm of section [ sec : algorithms ] can be used to explicitly construct the finitely generated intersection of two finitely generated subgroups using the perimeter techniques we have already introduced . the method is a 2-dimensional generalization of that described by stallings for graphs @xcite . the construction of the finitely generated intersection will proceed in two steps . the first step will be to reduce this property to a property of the fiber product of the spaces corresponding to these subgroups . the second step will be to show how this property can be achieved using perimeter reductions . let @xmath6 be a complex , let @xmath846 and @xmath847 be maps , and let @xmath848 be the _ diagonal _ of @xmath849 so @xmath850 . the _ fiber - product _ @xmath851 is defined to be the subspace of @xmath852 which is the preimage of @xmath387 in @xmath853 . identifying @xmath387 with @xmath6 , there is a natural map @xmath851 , and in fact , the following diagram commutes : @xmath854 notice that if both of the maps @xmath846 and @xmath847 have compact domain , then so does @xmath851 , and if both maps are immersions , then so is their fiber product . notice also that if @xmath58 is the map @xmath855 and @xmath96 is a subcomplex of @xmath6 , then @xmath856 is @xmath857 . note that the map @xmath858 induces a cell structure on @xmath856 such that the map @xmath858 is combinatorial . furthermore , if @xmath20 , @xmath96 , and @xmath6 are based spaces and the maps from @xmath20 and @xmath96 preserve basepoints , then @xmath859 and @xmath856 have natural basepoints . the _ based component _ @xmath13 of @xmath856 is the component of @xmath856 containing this basepoint . in particular we have @xmath860 . [ lem : fiberproduct ] let @xmath846 and @xmath847 be based maps and let @xmath120 , @xmath121 , and @xmath80 be their basepoints . if for every element of @xmath861 there is a closed path @xmath213 based at @xmath80 which lifts to closed paths @xmath862 and @xmath863 based at @xmath120 and @xmath121 , then the image of @xmath864 in @xmath14 is the intersection of the images of @xmath865 and @xmath826 . in particular , when @xmath20 and @xmath96 are compact , the fiber product @xmath856 will also be compact , and @xmath866 is finitely generated . since the based component of @xmath856 factors through @xmath20 and @xmath96 , the image of its fundamental group must be contained in the intersection of the images of their fundamental groups . on the other hand , by hypothesis , each element in this intersection has a representative which lifts both to @xmath20 and to @xmath96 , and thus to @xmath856 as well . in particular , @xmath866 also maps onto this intersection . while we have already proven in theorem [ thm : lqc - sc-1 ] that the following groups are locally quasiconvex and hence have the finitely generated intersection property , the following theorem gives an explicit and relatively efficient method of computing this intersection . [ thm : fgip ] let @xmath6 be a weighted @xmath536-@xmath537-complex @xmath538-@xmath539 $ ] . if every @xmath40-shell with @xmath543 @xmath541 $ ] is perimeter - reducing then @xmath14 has the finitely generated intersection property and the intersection of two finitely generated subgroups of @xmath14 can be constructed explicitly . first note that by theorem [ thm : fan - classification ] , @xmath6 satisifies the @xmath0-cell reduction hypothesis . let @xmath18 and @xmath660 be finitely generated subgroups of @xmath14 and let @xmath867 and @xmath868 be compact complexes chosen so that the images of their fundamental groups are @xmath18 and @xmath660 . we will now show by construction that @xmath869 is finitely generated . let @xmath870 and @xmath871 be obtained from @xmath872 and @xmath873 by running the perimeter reduction algorithm described in section [ sec : algorithms ] . when the algorithm stops , no further folds or @xmath0-cell perimeter reductions can be performed on @xmath870 or @xmath874 . thus , by the remark at the end of definition [ def : prh ] , @xmath875 and @xmath876 can now be viewed as subgroups of @xmath14 . next , let @xmath877 and @xmath878 be complexes obtained from @xmath879 and @xmath880 by attaching to every vertex @xmath374 , a copy of each @xmath0-cell in @xmath6 whose boundary cycle contains the image of @xmath374 in @xmath6 . these copies of @xmath0-cells are attached only at the vertex @xmath374 , and the maps into @xmath6 are extended in the obvious way . finally , let @xmath881 and @xmath882 be obtained by rerunning the perimeter reduction algorithm on @xmath883 and @xmath884 . note that each of these steps adds only a finite number of @xmath0-cells , and consequently , since @xmath872 and @xmath873 are compact , so are @xmath885 and @xmath886 . it should also be clear that @xmath887 and @xmath888 , and since running the perimeter reduction algorithm does not change the image of the fundamental group in @xmath6 , @xmath889 and @xmath890 . it only remains to show that @xmath891 and @xmath892 satisfy the conditions of lemma [ lem : fiberproduct ] . for each element of @xmath869 we can choose closed paths @xmath893 and @xmath894 whose images in @xmath6 represent this element . since the concatenation @xmath213 followed by the inverse of @xmath599 is null - homotopic in @xmath6 , there is a minimal area disc diagram @xmath895 whose boundary cycle is @xmath548 . let @xmath896 and @xmath897 be chosen ( among combinatorial paths homotopic to them ) so that the corresponding disc diagram @xmath387 is of minimal area . observe that neither @xmath898 nor @xmath899 contains a boundary arc which is the complement of at most @xmath1 pieces in @xmath387 ( @xmath0 pieces in the @xmath536-@xmath537 case ) . to see this , suppose that such a subpath existed in @xmath69 . by hypothesis , there would then be a perimeter reduction on @xmath213 which would yield a corresponding reduction of @xmath900 ( since no perimeter reductions exists for @xmath901 ) . but this would yield a new path @xmath69 and a new disc diagram @xmath387 with fewer @xmath0-cells . we therefore conclude by corollary [ cor : fan - classification ] , that @xmath387 is @xmath660-thin . it is now easy to see that @xmath213 actually lifts to @xmath886 as well as @xmath885 . thus @xmath891 and @xmath902 satisfy lemma [ lem : fiberproduct ] and the proof is complete . the fact that theorem [ thm : fgip ] uses the @xmath0-cell reduction hypothesis rather than the path reduction hypothesis is crucial . the examples below satisfy the path reduction hypothesis but fail to have the finitely generated intersection property . [ exmp : fgip - fails ] let @xmath6 be the standard @xmath0-complex of the presentation @xmath903 , [ b , t ] \rangle$ ] . if we assign a weight of @xmath21 to every side incident at the edge labeled @xmath904 and a weight of @xmath47 to every other side of the complex , the result is a @xmath536-@xmath537 complex which satisfies the path reduction hypothesis . however , the subgroup @xmath905 is not finitely generated . similarly , if we let @xmath6 be the standard @xmath0-complex of the presentation @xmath906 , and we assign a weight of @xmath47 to all of the sides incident at the edge labeled @xmath904 and a weight of @xmath21 to all of the other sides , the result is a @xmath549-@xmath550-complex which satisfies the path reduction hypothesis . since this group has a free factor which is commensurable with @xmath907 , it also fails to have howson s property . we note that the construction above works to explicitly compute the intersection between two subgroups of @xmath14 if @xmath6 satisfies the more general criterion of theorem [ thm : fan - diagramlqc ] . the main difference is that one attaches fans in @xmath590 instead of @xmath40-shells , and one adds a diameter @xmath714 `` neighborhood '' in passing from @xmath879 and @xmath880 to @xmath883 and @xmath884 . in this subsection we show how the intersections between specific subgroups can sometimes be shown to be finitely generated under a weaker set of assumptions . for example , even if we only assume the path reduction hypothesis , we can sometimes prove that certain subgroups @xmath908 have the property that @xmath909 is finitely generated for every finitely generated subgroup @xmath910 . [ thm : fgip - zero ] let @xmath6 be a compact based @xmath0-complex with non - negative integer weights assigned to the sides of its @xmath0-cells , and let @xmath682 be a based subgraph of @xmath437 with @xmath911 . if the weight of each @xmath0-cell in @xmath6 is positive and @xmath6 satisfies the path reduction hypothesis , then for any finitely generated subgroup @xmath18 in @xmath14 , the intersection @xmath912 is also finitely generated . let @xmath8 be the based covering space of @xmath6 corresponding to the subgroup @xmath18 . let @xmath913 be a based @xmath7-surjective subcomplex . according to theorem [ thm : gct ] , there exists a compact connected subcomplex @xmath914 such that @xmath915 , and such that @xmath3 is of minimal perimeter among all such complexes and hence such that @xmath916 is an isomorphism . let @xmath58 denote the map from @xmath4 to @xmath6 , and let @xmath13 denote the connected component of @xmath917 which contains the basepoint of @xmath4 . note that @xmath917 is the fiber product @xmath918 of the maps @xmath3 and @xmath919 . we will show that @xmath920 where @xmath13 is the based component of the fiber product . it is obvious that @xmath921 and so it is sufficient to show that @xmath922 . suppose that @xmath923 is a closed immersed path which represents an element of @xmath16 , then @xmath69 lifts to a closed path @xmath924 in @xmath8 . let @xmath925 and note that @xmath10 is compact and connected . since @xmath924 is a closed path , @xmath10 is formed from @xmath4 by adding finitely many arcs whose endpoints lie in @xmath4 . in particular , @xmath926 consists of a set of @xmath21-cells and perimeter zero @xmath47-cells , and so @xmath927 . if @xmath928 then @xmath188 is a proper free factor of @xmath929 and hence the map @xmath930 is not @xmath7-injective . part a of theorem [ thm : gct ] provides a complex @xmath330 with @xmath931 and @xmath932 . this contradicts the minimality of @xmath5 . we note that when @xmath6 satisfies the strict hypothesis , generators for an intersection can be computed . to see that the hypothesis that @xmath911 can not be dropped , let @xmath682 denote the @xmath935 subcomplex of example [ exmp : fgip - fails ] . the following application is an example from small - cancellation theory : in the group @xmath936 , the subgroup generated by @xmath120 and @xmath121 has a finitely generated intersection with every other finitely generated subgroup of @xmath19 . to apply theorem [ thm : fgip - zero ] we let @xmath6 be the standard @xmath0-complex of the presentation , and let @xmath682 be the graph consisting of the unique @xmath21-cell of @xmath6 together with the edges corresponding to the letters @xmath120 and @xmath121 . note that @xmath6 is a @xmath937-complex and that all its pieces are of length @xmath47 . also , if we assign a weight of @xmath21 to all of the sides present at the edges labeled @xmath120 and @xmath121 , and a weight of @xmath47 to all of the others , then the unique @xmath0-cell has weight @xmath559 , and the perimeter of each piece is at most @xmath1 . thus @xmath6 satisfies the path reduction hypothesis . recall that a magnus subgroup of @xmath19 is any subgroup which is generated by a proper subset @xmath940 . let @xmath6 be the standard @xmath0-complex of @xmath19 and let @xmath682 be the graph formed using the unique @xmath21-cell and the @xmath47-cells corresponding to the generators in @xmath96 . we assign a weight of @xmath21 to every side incident at an edge in @xmath682 and a weight of @xmath47 to all of the other sides . to show that @xmath6 , with this weighting , satisfies the path reduction hypothesis , by theorem [ thm : wt - one - rel ] we only need to show that the weight of the packet of the @xmath0-cell is at most the perimeter of a subword of ( a cyclic conjugate of ) @xmath23 . in order to apply theorem [ thm : wt - one - rel ] , the weight of @xmath50 must be strictly positive . this corresponds to the existence of a letter in @xmath23 which is not in @xmath682 . when this is not the case , the subgroup generated by @xmath96 is a free factor of @xmath19 and the result follows immediately from the theory of free products . thus , we may assume that the weight of @xmath23 is indeed positive , and that theorem [ thm : wt - one - rel ] applies . let @xmath498 be the number of times that elements outside of @xmath682 occur in @xmath23 . the weight of the unique @xmath0-cell is then @xmath941 , and the weight of its packet is @xmath942 . on the other hand , since the perimeter of a single edge in @xmath6 is at most @xmath941 , the perimeter of a subword of @xmath23 is at most @xmath943 . thus , the path reduction hypothesis is satisfied whenever @xmath797 . a subgroup of @xmath19 generated by elements @xmath944 is said to have _ solvable membership problem _ provided it is decidable whether an element @xmath945 lies in @xmath946 . if the membership problem is solvable for every finitely generated subgroup of @xmath19 , then @xmath19 is said to have a _ solvable generalized word problem_. the name alludes to the fact that it includes the question of membership in the trivial subgroup ( the word problem for @xmath19 ) as a particular case . [ lem : closing - paths ] let @xmath6 be a @xmath0-complex with non - negative integer weights assigned to the sides of its @xmath0-cells . if @xmath6 satisfies the ( weak ) path reduction hypothesis and @xmath140 is a map which does not admit a ( weak ) perimeter reduction or a fold , then every path @xmath947 whose image is a closed and null - homotopic path in @xmath6 will be a closed path in @xmath4 as well . let @xmath396 ( @xmath948 ) be a sequence of folds and perimeter reductions which starts with the path @xmath949 and ends with the trivial path . since @xmath140 does not admit any folds or ( weak ) perimeter reductions , all of the alterations to the path @xmath213 can be mimicked in the path @xmath613 . in particular , since the final path @xmath384 is the trivial path and since the endpoints of @xmath386 are the same throughout this process , the original path @xmath613 must have been closed . [ thm : generalizedwp ] let @xmath6 be a @xmath0-complex with non - negative integer weights assigned to the sides of its @xmath0-cells . if @xmath6 satisfies the path reduction hypothesis , then @xmath14 has a solvable generalized word problem . let @xmath950 be a set of closed paths in @xmath6 with a common basepoint . to decide whether the element of @xmath14 represented by @xmath477 is in the subgroup @xmath18 generated by the elements corresponding to the @xmath951 we proceed as follows . let @xmath316 be the wedge of @xmath82 closed paths and a single open path attached to the others at only one of its endpoints . we define the map @xmath952 so that it agrees with the maps @xmath953 and @xmath954 . let @xmath519 denote the basepoint of @xmath316 , and let @xmath526 be the other endpoint of the open path in @xmath316 ( see figure [ fig : membershipwedge ] ) . let @xmath330 be the final complex produced by running the perimeter reduction algorithm on the map @xmath952 . recall that the process also constructs a @xmath7-surjective map @xmath955 such that the composition @xmath956 is the original map @xmath952 . we claim that vertices @xmath519 and @xmath526 are identified under the map @xmath955 if and only if the element corresponding to the path @xmath957 lies in the subgroup @xmath18 . both directions of this implication need to be verified . first , observe that if @xmath519 and @xmath526 are identified in @xmath330 , then @xmath958 is sent to a closed path @xmath959 . since by the remark at the end of definition [ def : prh ] , @xmath332 can be considered a subgroup of @xmath14 whose image is the same as the image of @xmath319 ( which is @xmath18 ) , the element represented by @xmath477 must lie in the subgroup @xmath18 . conversely , suppose that the element represented by @xmath477 is indeed in the subgroup @xmath18 . then there exists a closed path @xmath960 whose image in @xmath6 is homotopic to @xmath477 . in particular , the path @xmath961 is a path in @xmath316 whose image in @xmath6 is null - homotopic . by lemma [ lem : closing - paths ] , the image of this path in @xmath330 must be closed . [ cor : generalizedwp ] if @xmath19 is a group which satisfies the hypotheses of theorem [ thm : wt - one - rel ] or theorem [ thm : coh - sc-1 ] , then each of its finitely generated subgroups has a decidable membership problem . gilbert baumslag . some problems on one - relator groups . in _ proceedings of the second international conference on the theory of groups ( australian nat . univ . , canberra , 1973 ) _ , pages 7581 . lecture notes in math . , vol . 372 . springer , berlin , 1974 .

a group is coherent if all its finitely generated subgroups are finitely presented . in this article we provide a criterion for positively determining the coherence of a group . this criterion is based upon the notion of the perimeter of a map between two finite @xmath0-complexes which is introduced here . in the groups to which this theory applies , a presentation for a finitely generated subgroup can be computed in quadratic time relative to the sum of the lengths of the generators . for many of these groups we can show in addition that they are locally quasiconvex . as an application of these results we prove that one - relator groups with sufficient torsion are coherent and locally quasiconvex and we give an alternative proof of the coherence and local quasiconvexity of certain @xmath1-manifold groups . the main application is to establish the coherence and local quasiconvexity of many small cancellation groups .

the jacobian elliptic function @xmath2 and the complete elliptic integral of the first kind @xmath3 play important roles in expressing exact solutions of , for example , the pendulum equation @xmath4 , a typical bistable equation @xmath5 , and so on . now we will propose new generalization of @xmath2 and @xmath3 . for constants @xmath6 and @xmath7 , we define a _ generalized jacobian elliptic function _ \to [ 0,1]$ ] with a modulus @xmath0 as @xmath9 where @xmath10 and @xmath11 we extend the domain of @xmath12 to @xmath13 so that we obtain a @xmath14-period function like the sine function , and call the extended function @xmath12 again . then , @xmath15 and @xmath16 when @xmath17 ; and @xmath18 and @xmath19 when @xmath20 , where @xmath21 is the generalized trigonometric function and @xmath22 is the half period of @xmath21 , which will be introduced in section [ sec : gjef ] below . therefore , @xmath12 is also generalization of both @xmath2 and @xmath21 . in the previous paper @xcite , the author proposed another generalization of @xmath2 and @xmath3 , and applied them to bifurcation problems for @xmath23-laplacian . as we will see in section [ sec : gjef ] , @xmath12 and @xmath24 above are defined in a slightly different way from those in @xcite , but @xmath12 also satisfies the following equation involving @xmath23-laplacian nevertheless . @xmath25 while the generalization of @xmath3 of @xcite converges to a finite value as @xmath26 when @xmath27 , the @xmath24 diverges to @xmath28 as @xmath26 for any @xmath29 . in this sense , @xmath12 has closer properties to @xmath2 than the function defined in @xcite . in the present paper , we will show the basis property of functions @xmath30 which means that the family of these functions is a basis in banach spaces . here , a sequence @xmath31 in a banach space @xmath32 is called a basis for @xmath32 if for every @xmath33 there exists a unique sequence of scalars @xmath34 such that @xmath35 in the strong sense . in general , when we try to find an approximation of a given function by a family of functions @xmath31 , it is desirable that @xmath31 is a basis which approximates to the function with convergence of higher order as possible . concerning this , for example , we have known an interesting study @xcite of boulton and lord . they study the best index @xmath36 for which @xmath37 approximates well to the solution of @xmath23-poisson problem , where @xmath38 and @xmath39 . the basis property is quite fundamental to such a stimulating problem . when @xmath17 , the sequence is the family of jacobian elliptic functions @xmath40 . in this case , craven @xcite proves that if the modulus @xmath0 satisfies @xmath41 , then the sequence is complete in @xmath42 . since the sequence is not orthogonal , we have no guarantee of its basis property . on the other hand , when @xmath20 , the sequence is the family of generalized trigonometric functions @xmath43 . on the sequence for @xmath44 , binding et al.@xcite first studied the basis property . recently , edmunds et al.@xcite show that if , for example , @xmath45 , then the sequence is a basis in @xmath46 for any @xmath47 . this paper deals with the basis property of for general @xmath6 and @xmath7 . our results involve results of @xcite when @xmath17 and @xcite when @xmath20 . [ thm : main ] let @xmath6 and @xmath48 . if @xmath49 then @xmath50 forms a riesz basis of @xmath42 and a schauder basis of @xmath51 for any @xmath47 when @xmath20 or @xmath52 where @xmath53 is the beta function and @xmath54 is a _ generalized inverse hyperbolic function _ @xmath55 the function @xmath56 is a monotone increasing function from @xmath57 onto @xmath58 . from theorem [ thm : main ] , we obtain an improvement of craven s result stated in remark of ( * ? ? ? * theorem 2 ) . [ cor : main ] if @xmath59 , then @xmath50 forms a riesz basis of @xmath42 and a schauder basis of @xmath51 for any @xmath47 when @xmath20 or @xmath60 in particular , the sequence of jacobian elliptic functions @xmath40 does so when @xmath41 . we will give another corollary of theorem [ thm : main ] , whose conditions are verified easier than and . [ thm : main ] let @xmath6 and @xmath48 . if @xmath61 then @xmath50 forms a riesz basis of @xmath42 and a schauder basis of @xmath51 for any @xmath47 when @xmath62^{\frac1q}.\ ] ] \(i ) if @xmath63 , i.e. , @xmath64 , then and hold . ( ii ) case @xmath20 in corresponds to the result of ( * ? ? ? * theorem 4.4 ) . ( iii ) when @xmath65 , the value of the right - hand side of is about @xmath66 , which is not so satisfactory as @xmath67 of corollary [ cor : main ] . however , we can check much easier than . the paper is organized as follows . in section [ sec : pb ] we give a summary of general properties of bases in banach spaces . in section [ sec : gjef ] we recall the generalized trigonometric functions and introduce new generalization of jacobian elliptic functions . in section [ sec : psk ] we observe properties of the generalized jacobian elliptic function @xmath12 and its quarter period @xmath24 . to show that the sequence is a basis in @xmath51 for any @xmath68 , we depend on the strategy of binding et al . @xcite and edmunds et al . our main device is a linear mapping @xmath69 of @xmath51 , satisfying @xmath70 , where @xmath71 , and decomposing into a linear combination of certain isometries . in section [ sec : tot ] we show that @xmath69 is a bounded operator for @xmath72 . section [ sec : bit ] is devoted to the proof of boundedness of the inverse for the ranges and . in this section we will give a summary of properties of bases in banach spaces . for details , we can refer to gohberg and kren @xcite , higgins @xcite , and singer @xcite . a sequence @xmath73 in an infinite dimensional banach space @xmath32 is called a _ basis of @xmath32 _ if for every @xmath74 there exists a unique sequence of scalars @xmath34 such that @xmath75 ( i.e. , such that @xmath76 ) . a basis @xmath73 of a topological linear space @xmath77 is said to be a _ schauder basis _ of @xmath77 , if all coefficient functionals @xmath78 , are continuous on @xmath77 . [ prop : schauder ] every basis of a banach space is a schauder basis of this space . see ( * ? ? ? * theorem 3.1 , p.20 ) . a basis @xmath73 of a banach space @xmath32 is said to be \(a ) a _ bessel basis _ , if @xmath79 i.e. , there exists a constant @xmath80 such that we have @xmath81 for all finite sequences of scalars @xmath82 . \(b ) a _ hilbert basis _ , if @xmath83 i.e. , there exists a constant @xmath84 such that we have @xmath85 for all finite sequences of scalars @xmath82 . \(c ) a _ riesz basis _ , if it is both a bessel basis and a hilbert basis , i.e. , there exist two constants @xmath80 and @xmath84 such that we have @xmath86 for all finite sequences of scalars @xmath82 . [ exmp : singer343 ] in the space @xmath87 , the sequence @xmath73 , where @xmath88,\ n=1,2,\ldots)\ ] ] is a bounded bessel basis if @xmath89 and a bounded hilbert basis if @xmath90 . in particular , it is a riesz basis if @xmath91 . see ( * ? ? ? * example 11.1 , pp.342345 ) . we call two sequences @xmath92 and @xmath93 in banach space @xmath32 _ equivalent _ if there exists a linear homeomorphism ( i.e. , bounded , linear and invertible operator ) @xmath69 on @xmath32 such that @xmath94 for every @xmath95 . note that by ` invertible ' we mean that @xmath96 exists and is bounded on all of @xmath32 . [ prop : higgins75 ] if a sequence @xmath93 in a banach space is equivalent to a basis @xmath92 , it too is a basis . see ( * ? ? ? * lemma , p.75 ) . furthermore , we have [ thm : bari311 ] let @xmath97 be a hilbert space . the following assertions are equivalent . @xmath98 the sequence @xmath92 forms a basis of @xmath97 , equivalent to an orthonormal basis . @xmath99 the sequence @xmath92 is complete and a riesz basis in @xmath97 . see the first and third assertions of ( * ? ? ? * theorem 2.1 , pp.310311 ) . this section is devoted to the definitions of two kinds of generalized functions . _ generalized trigonometric functions _ were introduced in 1879 by e.lundberg ( see lindqvist and peetre @xcite ) . after that , these functions have been developed mainly by elbert @xcite , lindqvist @xcite , drbek and mansevich @xcite , and lang and edmunds @xcite . for any constants @xmath6 , we define @xmath22 by @xmath100 where @xmath101 and @xmath102 are the beta and gamma functions , respectively . then , for any @xmath103 $ ] we define @xmath21 by @xmath104 clearly , @xmath21 is an increasing function in @xmath105 from @xmath106 $ ] onto @xmath107 $ ] . we extend the domain of @xmath21 to @xmath108 $ ] by @xmath109 , and furthermore , to the whole of @xmath13 by @xmath110 , so that @xmath21 has @xmath111-periodicity . we can see that @xmath112 and @xmath113 . moreover , the function @xmath114 satisfies that @xmath115 , and @xmath116 if @xmath117 . we agree that @xmath118 and @xmath119 denote @xmath120 and @xmath121 when @xmath44 , respectively . in that case , we can also refer to @xcite . using @xmath21 , for @xmath103 $ ] we also define @xmath122 clearly , @xmath123 is a decreasing function in @xmath105 from @xmath106 $ ] onto @xmath107 $ ] . we extend the domain of @xmath123 to @xmath124 $ ] by @xmath125 , and furthermore , to the whole of @xmath13 in the same way as @xmath21 . then , @xmath123 has @xmath111-periodicity . we can see that @xmath126 . an analogue of @xmath127 is obtained by defining @xmath128 for those values of @xmath105 at which @xmath129 . this means that @xmath130 is defined for all @xmath131 except for the points @xmath132 . we denote by @xmath133 and @xmath134 as for the case @xmath119 . the functions @xmath119 and @xmath135 are useful for prfer transformation of half - linear differential equations . for this , see @xcite . these functions satisfy , for @xmath136 , @xmath137 the case @xmath138 for some @xmath139 is as follows . @xmath140 it is useful to collect formulae for case @xmath141 and @xmath142 for some @xmath139 . @xmath143 in particular , it is important that for any @xmath6 @xmath144 we can find many other properties of these functions in @xcite . there are some different definitions of @xmath123 from . for example , drbek and mansevich @xcite define @xmath123 by @xmath145 and so gives @xmath146 which is slightly different from . the fact that @xmath21 satisfies is essential , independently of the definition of @xmath123 . [ lem : pipp ] the number @xmath147 is an increasing function in @xmath139 and @xmath148 and @xmath149 . putting @xmath150 in the definition of @xmath147 , we have @xmath151 it suffices to show that @xmath152 is decreasing on @xmath57 . using a formula of euler ( see example 36 in @xcite ) : @xmath153 we have @xmath154 clearly , the right - hand side is decreasing in @xmath155 ( note that @xmath156 ) , so that @xmath152 is also decreasing on @xmath57 . furthermore , since @xmath157 we obtain the values of limits . we can find another proof of proposition [ lem : pipp ] in ( * ? ? ? * lemma 2.4 ) , in which they use the fact that the area of @xmath158-circle @xmath159 is @xmath147 ( see also @xcite ) . in the fashion of the classical jacobian elliptic functions , we define new transcendental functions . let @xmath6 . for any @xmath7 we define @xmath24 by @xmath160 then , for any @xmath7 and @xmath161 $ ] we define @xmath162 by @xmath163 clearly , @xmath162 is an increasing function in @xmath105 from @xmath164 $ ] onto @xmath107 $ ] . we extend the domain of @xmath162 to @xmath165 $ ] by @xmath166 , and furthermore , to @xmath13 by @xmath167 , so that @xmath162 has @xmath14-periodicity . we can see that @xmath168 , and @xmath169 . moreover , the function @xmath170 satisfies that @xmath115 , and @xmath116 if @xmath117 . using @xmath162 , for @xmath161 $ ] we also define @xmath171 clearly , @xmath172 and @xmath173 are decreasing functions in @xmath105 from @xmath164 $ ] onto @xmath107 $ ] . we extend the domains of @xmath172 and @xmath173 to @xmath174 $ ] in the same way of @xmath123 , and furthermore , to @xmath13 by @xmath175 and @xmath176 , respectively . this implies that @xmath172 and @xmath173 have @xmath14- and @xmath177-periodicity . we can see that @xmath178 and @xmath179 . these functions satisfy , for @xmath180 @xmath181 in case @xmath138 for some @xmath139 , we write @xmath182 and @xmath183 by @xmath184 and @xmath185 , respectively . for the case , the formulae above become as follows . @xmath186 we also state the case @xmath141 and @xmath142 for some @xmath139 . @xmath187 moreover , @xmath170 satisfies @xmath25 as mentioned in introduction , the author @xcite has introduced another generalized jacobian elliptic functions , which also include both the jacobian elliptic functions and the generalized trigonometric functions . however , we should note that the definitions above of @xmath24 and @xmath12 are slightly different from those of @xcite , in which the common index of @xmath188 to and is not @xmath189 but @xmath190 . on account of the index , @xmath24 has similar asymptotic behavior near @xmath191 as @xmath3 , indeed , @xmath192 for any @xmath6 . to observe the convergence properties of generalized jacobian elliptic functions as @xmath26 , we will prepare _ generalized hyperbolic functions _ , for which similar definitions are seen in @xcite . for @xmath193 , we define @xmath194 by @xmath195 and extend its domain to @xmath13 by @xmath196 . using @xmath194 , for @xmath193 , we define @xmath197 and extend its domain to @xmath13 by @xmath198 . the function @xmath199 is defined by @xmath200 we agree that @xmath201 and @xmath202 denote @xmath203 and @xmath204 when @xmath44 , respectively . putting @xmath44 and @xmath205 in , we have @xmath206 then , it is easy to prove the following properties : for any @xmath6 and all @xmath131 , @xmath207 in this section we observe some properties of generalized jacobian elliptic function @xmath12 and its quarter period @xmath24 . the function @xmath208 satisfies that @xmath209 , @xmath210 for @xmath180 , @xmath211 $ ] , and @xmath212 if @xmath117 , then @xmath213 $ ] and @xmath214 when @xmath27 , we see that @xmath215 and @xmath216_0^{k_{pq}(k)}=1.\ ] ] to obtain the estimate of @xmath217 in lemma [ lem : estimatek ] below , we state tchebycheff s integral inequality in @xcite . [ lem : tchebycheff ] let @xmath218 \to \mathbb{r}$ ] be integrable functions , both increasing or both decreasing . furthermore , let @xmath219 \to \mathbb{r}$ ] be a positive , integrable function . then @xmath220 if one of the functions @xmath221 or @xmath222 is nonincreasing and the other nondecreasing , then the inequality in is reversed . concerning the following lemmas [ lem : monotonicity][lem : homo ] , we can refer to @xcite for the corresponding results of @xmath21 and @xmath22 , which are in case @xmath20 of @xmath12 and @xmath24 . lemma [ lem : estimatek ] extends an estimate of @xmath3 by qi and huang ( * ? ? ? * eq.(10 ) ) to that of @xmath217 for any @xmath139 . since @xmath223 , it suffices to observe properties of @xmath224 in order to study those of . [ lem : monotonicity ] for each @xmath225 $ ] , @xmath226 is decreasing on @xmath58 for any fixed @xmath227 . @xmath228 is decreasing on @xmath58 for any fixed @xmath229 . @xmath230 is increasing on @xmath231 for any fixed @xmath6 . first we will show that @xmath232 is decreasing in @xmath72 . let @xmath233 . putting @xmath234,\ ] ] we have @xmath235 where @xmath236 and @xmath237 since @xmath238 is increasing in @xmath239 when @xmath240 , it is easy to see that @xmath241 which means that @xmath242 , i.e , @xmath243 for each @xmath244 $ ] . thus , @xmath245,\ ] ] namely , @xmath246.\end{gathered}\ ] ] therefore we conclude that @xmath247,\ ] ] so that @xmath232 is decreasing in @xmath29 . the assertions for @xmath36 and @xmath0 are proved in a similar way . it is enough to replace @xmath248 by @xmath249 respectively , and to replace @xmath238 by @xmath250 respectively . [ lem : convexity ] let @xmath6 and @xmath7 . then , for any @xmath251 $ ] , @xmath252 let @xmath208 . putting @xmath253 in , we have @xmath254 so that @xmath255 since @xmath256 , we obtain @xmath257 which implies the assertion . [ lem : kpq ] @xmath258 is decreasing on @xmath58 for any fixed @xmath227 . @xmath259 is decreasing on @xmath58 for any fixed @xmath229 . @xmath260 is increasing on @xmath231 and @xmath261 for any fixed @xmath6 . the assertion on @xmath23 immediately follows from @xmath262 the remaining parts also follow from the form of @xmath24 . [ lem : symmetry ] let @xmath6 and @xmath7 . then @xmath263 putting @xmath264 in , we have @xmath265 where @xmath266 . furthermore , setting @xmath267 , we obtain @xmath268 the integration of the right - hand side is equal to @xmath269 . [ lem : homo ] let @xmath6 and @xmath7 . then @xmath270 where @xmath48 . case @xmath63 follows from only lemma [ lem : monotonicity ] . case @xmath271 is also proved similarly after using lemma [ lem : symmetry ] . [ lem : estimatek ] let @xmath139 and @xmath272 . then @xmath273 putting @xmath274 in the definition of @xmath217 and using , we have @xmath275 we follow the proof of qi and huang @xcite , which deals with the case @xmath276 . let @xmath277 or @xmath278=[0,\pi_{r'r}/2]$ ] in tchebycheff s integral inequality of lemma [ lem : tchebycheff ] , then we obtain @xmath279 or @xmath280 by and , easy calculation gives @xmath281 moreover , putting @xmath282 , we have @xmath283 similarly , putting @xmath284 and @xmath285 , we obtain @xmath286 thus , we accomplished the first and second inequalities of . finally , from the equality above , we obtain the third inequality of as @xmath287 the graphs of terms of for @xmath276 can be shown in figure [ fig : kpq ] . . the black line and the gray lines indicate @xmath217 and the others , respectively.,width=188 ] let @xmath47 be an arbitrary number . in this section , we will make the functions @xmath288 correspond to the sine series @xmath289 which form a basis of @xmath51 . [ thm : homeomorphism ] if there is a linear homeomorphism @xmath69 of @xmath51 satisfying @xmath290 , then @xmath50 forms a riesz basis of @xmath42 and a schauder basis of @xmath51 for any @xmath47 . by example [ exmp : singer343 ] , any function in @xmath291 has a unique sine - cosine series representation . for any @xmath292 , we can thus represent its odd extension to @xmath291 uniquely in a sine series , so the @xmath293 form a basis of @xmath51 . since @xmath294 and @xmath295 are equivalent , according to proposition [ prop : higgins75 ] the same is true for the @xmath296 . it follows from proposition [ prop : schauder ] that they form a schauder basis of @xmath51 . the argument for a riesz basis when @xmath297 is similar and follows from proposition [ thm : bari311 ] . in the remainder of this section we define @xmath69 as a linear combination of certain isometries of @xmath51 . then we show that @xmath69 is a bounded operator satisfying @xmath290 , for all @xmath6 . the functions @xmath296 have fourier sine series expansions @xmath298 where @xmath299 an argument involving symmetry with respect to the middlepoint @xmath300 easily shows that @xmath301 whenever @xmath302 is even . on account of this property , we can show @xmath303 by using @xmath304 as follows . @xmath305 in what follows we will often denote @xmath306 by @xmath307 . we first find a bound on @xmath308 which will be crucial in the definition of @xmath69 below . since @xmath309 if @xmath310 is even , we may assume that @xmath310 is odd . integration by parts ensures that @xmath311 where the integrals exist because @xmath312 . in fact shows that @xmath313 in order to construct the linear operator @xmath69 , we next define isometries @xmath314 of the banach space @xmath51 by @xmath315 , where @xmath316 is its successive antiperiodic extension of @xmath222 over @xmath317 by @xmath318 on @xmath107 $ ] , and @xmath319 notice that @xmath320 . [ lem:7 ] the maps @xmath314 are isometric linear transformations of @xmath51 for all @xmath321 and @xmath47 . @xmath322 we now define @xmath323 by @xmath324 lemma [ lem:7 ] , the triangle inequality and ensure that @xmath69 is a bounded everywhere - defined operator with @xmath325 . we conclude this section by showing that @xmath70 . indeed , by virtue of , @xmath326 in this section we complete the proof of theorem [ thm : main ] by showing that @xmath69 has a bounded inverse for all @xmath327 and @xmath0 satisfying and . observe that and the triangle inequality give @xmath328 so that lemma [ lem:7 ] and the fact @xmath329 ( the identity operator on @xmath51 ) give @xmath330 thus , by c.neumanns theorem ( * ? ? ? * theorem 2 in p.69 ) , to show that @xmath69 has a bounded inverse , it is sufficient to prove @xmath331 assume and let @xmath0 be a number satisfying . from lemmas [ lem : homo ] and [ lem : estimatek ] , we have @xmath336 using to the right - hand side , we obtain @xmath337 thus , and give @xmath338 which implies of theorem [ thm : general ] . we have thus proved the theorem . suppose that @xmath36 and @xmath158 satisfy . since @xmath340 as the proof of corollary [ cor : main ] , we obtain @xmath345 which implies . furthermore , if we assume , then in lemma [ lem : estimatek ] shows @xmath346 so that holds , and theorem [ thm : main ] gives corollary [ thm : main ] . 99 p.binding , l.boulton , j.epika , p.drbek and p.girg , basis properties of eigenfunctions of the p - laplacian , _ _ proc.amer.math.soc.__**134 * * ( 2006 ) , no.12 , 34873494 . p.j.bushell and d.e.edmunds , remarks on generalized trigonometric functions , _ _ rocky mountain j.math.__**42 * * ( 2012 ) , no.1 , 2557 . l.boulton and g.lord , approximation properties of the @xmath36-sine bases , _ proc.r.soc.lond.ser.a math.phys.eng.sci._ * 467 * ( 2011 ) , no.2133 , 26902711 . b.d.craven , stone s theorem and completeness of orthogonal systems , _ j.austral.math.sci . _ * 12 * ( 1971 ) , 211 - 223 . m.del pino , m.elgueta , r.mansevich , a homotopic deformation along @xmath23 of a leray - schauder degree result and existence for @xmath347 , _ j. differential equations _ * 80 * ( 1989 ) , 113 . o.dol , half - linear differential equations , _ handbook of differential equations _ , 161357 , elsevier / north - holland , amsterdam , 2004 . o.dol and p.ehk , _ half - linear differential equations _ , north - holland mathematics studies , 202 . elsevier science b.v . , amsterdam , 2005 . p.drbek , p.krej and p.tak , _ nonlinear differential equations _ , papers from the seminar on differential equations held in chvalatice , june 29july 3 , 1998 . chapman & hall / crc research notes in mathematics , 404 . chapman & hall / crc , boca raton , fl , 1999 . p.drbek and r.mansevich , on the closed solution to some nonhomogeneous eigenvalue problems with @xmath23-laplacian , _ differential integral equations _ * 12 * ( 1999 ) , 773788 . d.e.edmunds , p.gurka and j.lang , properties of generalized trigonometric functions , _ j.approx.theory _ * 164 * ( 2012 ) , no.1 , 4756 . a.elbert , a half - linear second order differential equation . _ qualitative theory of differential equations , vol . ( szeged , 1979 ) , pp . 153180 , colloq . janos bolyai , 30 , north - holland , amsterdam - new york , 1981 . i.c.gohberg and m.g.kren , _ introduction to the theory of linear nonselfadjoint operators _ , translated from the russian by a. feinstein . translations of mathematical monographs , vol . 18 american mathematical society , providence , r.i . 1969 j.r.higgins , _ completeness and basis properties of sets of special functions _ , cambridge tracts in mathematics , vol . cambridge university press , cambridge - new york - melbourne , 1977 . j.lang and d.e.edmunds , _ eigenvalues , embeddings and generalised trigonometric functions _ , lecture notes in mathematics , 2016 . springer , heidelberg , 2011 . p.lindqvist , some remarkable sine and cosine functions , _ _ ricerche mat.__**44 * * ( 1995 ) , no.2 , 269290 . p.lindqvist and j.peetre , @xmath23-arclength of the @xmath36-circle , _ the mathematics student _ * 72 * ( 2003 ) , 139145 . p.lindqvist and j.peetre , _ comments on erik lundberg s 1879 thesis . especially on the work of gran dillner and his influence on lundberg _ , memorie dellinstituto lombardo ( classe di scienze matem.nat . ) 31 , 2004 . d.s.mitrinovi , _ analytic inequalities _ , in cooperation with p. m. vasi , die grundlehren der mathematischen wissenschaften , band 165 springer - verlag , new york - berlin , 1970 . y.naito , uniqueness of positive solutions of quasilinear differential equations , _ differential integral equations _ * 8 * ( 1995 ) , no . 7 , 1813 - 1822 . f.qi and z.huang , inequalities of the complete elliptic integrals , tamkang j.math.*29 * ( 1998 ) , no.3 , 165169 . i.singer , _ bases in banach spaces . , die grundlehren der mathematischen wissenschaften , band 154 . springer - verlag , new york - berlin , 1970 . s.takeuchi , generalized jacobian elliptic functions and their application to bifurcation problems associated with @xmath23-laplacian , _ _ j.math.anal.appl.__**385 * * ( 2012 ) , no.1 , 2435 . e.t.whittaker and g.n.watson , _ a course of modern analysis _ , an introduction to the general theory of infinite processes and of analytic functions ; with an account of the principal transcendental functions . reprint of the fourth ( 1927 ) edition . cambridge mathematical library . cambridge university press , cambridge , 1996 . k.yosida , _ functional analysis _ , reprint of the sixth ( 1980 ) edition . classics in mathematics . springer - verlag , berlin , 1995 .

the jacobian elliptic functions are generalized to functions including the generalized trigonometric functions . the paper deals with the basis property of the sequence of generalized jacobian elliptic functions in any lebesgue space . in particular , it is shown that the sequence of the classical jacobian elliptic functions is a basis in any lebesgue space if the modulus @xmath0 satisfies @xmath1 .

an important feature of the solar convection zone is the presence of differential rotation in the form of a decrease in angular velocity from equator to the pole . this has been observed both in the surface layers ( e.g. snodgrass 1984 ) and deeper in the convection zone , as inferred from helioseismic measurements ( e.g. thompson _ et al . _ 1996 ) . furthermore , the differential rotation on the surface has been observed to vary with time ( e.g. howard & labonte 1980 ; snodgrass , howard , & webster 1985 ) . these so called torsional oscillations , which have periods of about 11 years , manifest themselves in the form of four alternating latitudinal bands of slightly faster and slower than average zonal flows which migrate towards the equator in about 22 years . these oscillations have also been confirmed by the analysis of the helioseismic data from the solar and heliospheric observatory ( soho ) spacecraft for the present solar cycle ( kosovichev & schou 1997 ; schou _ et al . _ 1998 ) . recent analysis of the helioseismic data , from both the michelson doppler imager ( mdi ) instrument on board the soho spacecraft and the global oscillation network group ( gong ) project , has also produced evidence that this banding signature is not merely a surface feature , but extends into the convection zone , to a depth of at least 8 percent in radius ( howe _ et al . these authors present data on departures of the reconstructed rotation rate from its temporal averages - the residuals - as a function of latitude at several target depths , which behave in a manner similar to the migration of sunspots during the solar cycle ( the ` butterfly diagram ' ) . this finding is also supported by the analysis of antia & basu ( 2000 ) , who use different data sets from gong and independent inversion techniques . the time - base of these observations is only a few years , less than a complete solar cycle . these torsional oscillations are thought to be produced as a consequence of the nonlinear interactions between the magnetic fields and the solar differential rotation . a zero order ` mean ' differential rotation is assumed to be maintained by the reynolds stresses of the turbulence ; this can be included as a constant ` background ' angular velocity , or explicitly parametrized , e.g. by the so called @xmath0effect ( e.g. rdiger 1989 ) . attempts have been made to explain the surface oscillations in terms of the effects of the lorentz force exerted by the large scale magnetic field on the azimuthal velocity field ( e.g. brandenburg & tuominen 1988 ) , or as a consequence of the ` quenching ' of the turbulence - dependent quantities by the magnetic field ( kitchatinov _ et al . _ 1999 ) . here we study this nonlinear coupling in the context of a two dimensional axisymmetric mean field dynamo model , in a spherical shell , in which the only nonlinearity is the action of the azimuthal component of the lorentz force of the dynamo generated magnetic field on the solar angular velocity . obtaining torsional oscillations in this way is also of interest in view of the fact that the lorentz force is of second order in the magnetic field , thus naturally leading to the excitation of hydrodynamical oscillations of about half the period of the magnetic oscillations . in the next section we introduce our model . section 3 contains our results and section 4 gives a brief discussion . here we shall assume that the gross features of the large scale solar magnetic field can be described by a mean field dynamo model , with the standard equation @xmath1 where @xmath2 , the term proportional to @xmath3 represents the effects of turbulent diamagnetism , and where the velocity field is taken to be of the form @xmath4 where @xmath5 , @xmath6 is a prescribed underlying rotation law and the component @xmath7 satisfies @xmath8 where @xmath9 is the operator @xmath10 and @xmath11 is the induction constant . the assumption of axisymmetry allows the field @xmath12 to be split into toroidal and poloidal parts , @xmath13 , and results in eq . ( [ mfe ] ) yielding two scalar equations for @xmath14 and @xmath15 . nondimensionalizing in terms of the solar radius @xmath16 and time @xmath17 , where @xmath18 is the maximum value of @xmath19 and letting @xmath20 , @xmath21 , @xmath22 , @xmath23 and @xmath24 , results in a system of equations for @xmath25 and @xmath7 , with the dynamo parameters @xmath26 , @xmath27 , @xmath28 , and @xmath29 , where @xmath30 is the solar surface equatorial angular velocity ( see moss & brooke 2000 for details ) . here @xmath31 and @xmath18 are the turbulent magnetic diffusivity and viscosity respectively and @xmath32 is the turbulent prandtl number . the density @xmath33 is assumed to be uniform , and stress free boundary conditions ensure angular momentum conservation . these equations were solved using the code and boundary conditions described in moss & brooke ( 2000 ) , over the range @xmath34 , @xmath35 , with uniform spacing in both @xmath36 and @xmath37 . the computational domain is the region @xmath38 ; with the solar convection zone proper being thought to occupy the region @xmath39 , the region @xmath40 can be thought of as an overshoot region / tachocline . in the following simulations we used a mesh resolution of 61 points uniformly distributed radially and 101 points uniformly distributed latitudinally ( over @xmath41 ) , but test runs were carried out at higher spatial resolutions . in this investigation , we took @xmath6 to be given in @xmath42 by an interpolation on the mdi data obtained from 1996 to 1999 ( howe _ et al . _ 2000 ) , depicted in fig . 1 . for @xmath43 we took @xmath44 , where @xmath45 ( cf . rdiger & brandenburg 1995 ) and @xmath46 with cubic interpolation to zero at @xmath47 and @xmath48 , with the convention that @xmath49 and @xmath50 . also , in order to take into account the likely decrease in the turbulent diffusion coefficient @xmath19 in the overshoot region , we allowed a simple linear decrease from @xmath51 at @xmath52 to @xmath53 in @xmath54 . we monitor the time evolution of the total magnetic energy @xmath55 and the global parity of the magnetic field , defined as @xmath56 , where @xmath57 and @xmath14 refer respectively to the parts of the magnetic field that have symmetry or antisymmetry with respect to the equatorial plane ( see also brandenburg _ et al . _ 1989 ) . thus @xmath58 and @xmath59 correspond to symmetric and antisymmetric fields respectively . we calibrated our model so that near marginal excitation the cycle period was about 22 years . this determined @xmath60 , corresponding to @xmath61 @xmath62s@xmath63 , given the known values of @xmath30 and @xmath16 . the first solutions to be excited in the linear theory are odd parity ( @xmath64 ) limit cycles , which in this case have marginal dynamo number @xmath65 . the even parity ( @xmath58 ) solutions are also excited at similar marginal dynamo numbers of @xmath66 . we considered two values of the prandtl numbers , @xmath67 and @xmath68 . # 1#20.5#1 with these parameter values , we found that this model , with underlying zero order angular velocity chosen to be consistent with the recent ( mdi ) helioseismic data ( fig . [ mdi ] ) , is capable of producing butterfly diagrams which are in qualitative agreement with the observations . an example of this is depicted in fig . [ calph=-3.2.pr=1.0.mdi2.toroidal.butterfly ] . the polar feature is rather too strong we have checked that this can be rectified by modifying slightly the spatial dependence of @xmath43 , by for example choosing @xmath69 . # 1#20.42#1 we also found that this model successfully produced torsional oscillations in the convection zone , similar to those deduced from recent helioseismic data . to compare our model with these results , we have plotted in figs . [ calph=-3.2.pr=0.1.mdi2.velocity ] and [ calph=-3.2.pr=1.0.mdi2.velocity ] the variations of the rotation rate with latitude and time ( ` butterfly diagrams ' ) to reveal the migrating banded zonal flows . in these models , the basic magnetic field parity is odd ( @xmath64 ) . # 1#20.42#1 # 1#20.42#1 # 1#20.42#1 # 1#20.42#1 for the sake of comparison with observational results , we have also plotted in fig . [ calph=-3.2.pr=1.0.mdi2.time.series.residuals ] the evolution of the residual rotation rate with time , at radius @xmath70 and latitude @xmath71 degrees . 1#20.40#1 as can be seen , consistent with the observations , in each hemisphere there are alternating latitudinal bands , with the width of approximately 10 degrees , of slightly faster and slower than average zonal flows . these migrate towards the equator in about 22 years , and extend deep into the convection zone . the amplitudes of these oscillations increase with depth below the surface and depend on the parameters of our model , in particular the prandtl number . for @xmath68 , these amplitudes range from about 0.07 nhz at the surface to more than @xmath72 nhz towards the bottom of the convection zone , somewhat lower than , but in principle compatible with , the results of howe _ the torsional oscillations present in our model have periods half that of the period of the global magnetic field which is compatible with the observed period of the oscillations at the surface and consistent with the observed behaviour inside the convection zone . the strictly odd parity models presented have an equatorial feature in the ` butterfly diagrams ' for the velocity perturbations which does not appear to be present in the current inversions of the observational data . this becomes weaker as prandtl number increases . we also note in passing that these figures are almost identical for cases where even parity solutions are found , except that this feature is then absent . we further point out that the large scale solar magnetic field is probably of _ mixed _ global parity ( predominantly odd ) see , e.g. , pulkinnen _ et al . _ we have studied a solar dynamo model , calibrated to have the correct cycle period , with a mean rotation law given by recent helioseismic observations . this model produces butterfly diagrams in qualitative agreements with those observed . we have shown that a nonlinear coupling between the magnetic field and the solar differential rotation , where the nonlinearity is due to the action of the azimuthal component of the lorentz force , is capable of producing torsional oscillations , with period of about 11 years , which penetrate into the convection zone and which migrate towards the equator in about 22 years . the period of these oscillations is about half that of the period of the global magnetic fields . this is in agreement with the observed period of the torsional oscillations at the surface . for the oscillations inside the convection zone , this is a testable prediction , not contradicted by the current helioseismic observations , which so far extend over an interval less than a solar cycle . we note also that solutions with even parity ( @xmath73 ) show slightly larger amplitudes without the equatorial feature . the current inversions of the helioseismological data seem to suggest that the torsional oscillations largely disappear below about @xmath74 , in contrast to our model oscillations . however , there are uncertainties in these inversions , specially at the deeper levels . at the same time our dynamical model is oversimplified and substantial improvements can be made . in particular , the predicted amplitudes for torsional oscillations in our model are likely to be affected by our assumption of uniform density in eq . ( [ ns ] ) . nevertheless , we find it interesting that some of the major features in the torsional oscillations can be readily reproduced . a more detailed study of these oscillations , including more realistic density profiles , is in progress . we would like to thank rachel howe and mike thompson for many useful discussions regarding the helioseismic observations and for providing us with the data displayed in fig . ec is supported by a pparc fellowship and rt benefited from pparc uk grant no .

recent analysis of the helioseismic observations indicate that the previously observed surface torsional oscillations extend significantly downwards into the solar convection zone . in an attempt to understand these oscillations , we study the nonlinear coupling between the magnetic field and the solar differential rotation in the context of a mean field dynamo model , in which the nonlinearity is due to the action of the azimuthal component of the lorentz force of the dynamo generated magnetic field on the solar angular velocity . the underlying zero order angular velocity is chosen to be consistent with the most recent helioseismic data . the model produces butterfly diagrams which are in qualitative agreement with the observations . it displays torsional oscillations that penetrate into the convection zone , and which with time migrate towards the equator . the period of these oscillations is found to be half that of the period of the global magnetic fields . this is compatible with the observed period of the surface torsional oscillations . inside the convection zone , this is a testable prediction that is not ruled out by the observations so far available .

the formation and condensation of excitonic bound states of conduction - band electrons and valence - band holes belongs surely to one of the most exciting ideas of contemporary solid state physics . although the excitonic condensation has been predicted a long time ago @xcite , no conclusive experimental proof of its existence has been achieved yet . the latest experimental studies of materials with strong electronic correlations showed however , that there are a few promising candidates for the experimental verification of the excitonic condensation . the first one is the mixed valence compound @xmath4 , where detailed studies of the pressure - induced semiconductor - semimetal transition indicate that excitons are created in a large number and condense below 20 k @xcite . moreover , the charge - density - wave ( cdw ) state observed in the layered transition - metal dichalcogenide @xmath5 was claimed to be of excitonic origin @xcite . quite recently , as a further candidate for the excitonic state , a quasi one - dimensional system @xmath6 has raised and attracted much experimental attention @xcite . in this material the flat band was observed by the arpes experiment , which was interpreted to be due to excitonic condensation . these results have stimulated further experimental and theoretical studies with regard to the formation and possible condensation of excitonic bound states of electron and holes in correlated systems . at present , it is generally accepted that the minimal theoretical model for a description of excitonic correlations in these material could be the falicov - kimball model @xcite and its extensions @xcite . in its original form , the falicov - kimball model describes a two - band system of the itinerant @xmath7 electrons ( with the nearest - neighbor @xmath7-electron hopping constant @xmath8 ) and the localized @xmath0 electrons that interact only via a local @xmath0-@xmath7 coulomb interaction @xmath9 : @xmath10 where @xmath11 and @xmath12 are the creation and annihilation operators of spinless electrons in the @xmath13 orbital at site @xmath14 and @xmath15 is the position of the @xmath0-level energy . in what follows we consider @xmath16 and all energies are measured in units of @xmath8 . since the local @xmath0-electron number @xmath17 is strictly conserved quantity , the @xmath7-@xmath0 electron coherence can not be established in this model . this shortcoming can be overcome by including an explicit local hybridization @xmath18 between the @xmath7 and @xmath0 orbitals . however , the hybridization between @xmath7 and @xmath0 orbitals is not the only way to develop @xmath7-@xmath0 coherence . analytical and numerical studies of batista et al . @xcite showed that a finite @xmath0-electron bandwidth @xmath19 also induces @xmath7-@xmath0 coherence , and thus it can lead to an excitonic condensate even in the absence of direct @xmath7-@xmath0 hybridization . later this model has been used extensively to describe different phases in the ground state and specially properties of the excitonic phase @xcite . it was found that the ground state phase diagram exhibits very simple structure consisting of only four phases , and namely , the full @xmath7 and @xmath0 band insulator , the excitonic insulator , the cdw and the staggered orbital order . the excitonic insulator is characterized by a nonvanishing @xmath20 average . the cdw is described by a periodic modulation in the total electron density of both @xmath0 and @xmath7 electrons , and the staggered orbital order is characterized by a periodic modulation in the difference between the @xmath0 and @xmath7 electron densities . an extension of this model with local hybridization between the @xmath0 and @xmath7 orbitals has been studied very recently in our work @xcite . the numerical analysis of the excitonic momentum distribution @xmath21 ( with @xmath22 , where l denotes the number of lattice sites ) showed that this quantity diverges for @xmath23 , signalizing a bose - einstein condensation of preformed excitons . moreover , it was found that the density of zero momentum excitons @xmath24 ( as well as the total exciton density @xmath25 ) depends strongly on the values of the coulomb interaction @xmath9 and that already for relatively small values of @xmath9 ( @xmath26 ) the significant fraction of @xmath27 excitons is in the zero - momentum state . in the current paper we examine in detail how robust is the excitonic state against the changes of other model parameters . in particular , we consider effects of the non - zero f - electron hopping @xmath28 , the non - local hybridization @xmath29 and the increasing dimension of the system , which together , but also separately model more realistically the physical situation in real rare - earth compounds than the model considered in our previous paper ( @xmath30 , @xmath31 ) . the crucial role of these factors in a correct description of ground state properties of these materials has been already confirmed by previous analytical and numerical studies . indeed , it was found , that the f - band ( f - electron hopping ) of the opposite parity than the d - band , stabilizes the excitonic phase , but only in the dimension @xmath32 @xcite , while the nonlocal hybridization ( in the mean - field approximation ) fully destroys the excitonic phase @xcite . it should be noted that the comfirmation ( refutation ) of the last result by exact calculations is of crucial importance for a correct description of ground states properties of some rare earth systems and especially the mixed valence compounds , like @xmath33 @xcite . in these materials the local hybridization is forbidden and only the nonlocal one ( with inversion symmetry ) between the nearest - neighbour f and d orbitals , is allowed @xcite . on the base of the above mentioned facts we expect the strong effects of these factors also on the formation and condensation of the excitonic bound states . besides these factors , we examine also the influence of the f - level position on the density of @xmath34 excitons , since the simple parametrization between the f - level position and external pressure @xmath35 , widely used in the literature @xcite allows us to predict , at least qualitatively , the behaviour of this quantity with the external pressure . to examine effects of the above mentioned factors on the formation and condensation of excitonic bound states in the generalized falicov - kimball model we have performed exhaustive dmrg studies of the model hamiltonian @xmath36 for a wide range of the model parameters at the total electron density @xmath37 . in all examined cases we typically keep up to 500 states per block , although in the numerically more difficult cases , where the dmrg results converge slower , we keep up to 1000 states . truncation errors @xcite , given by the sum of the density matrix eigenvalues of the discarded states , vary from @xmath38 in the worse cases to zero in the best cases . and the total exciton density @xmath39 as functions of local hybridization @xmath40 calculated by dmrg method for three different values of the interband coulomb interaction @xmath9 ( @xmath41 ) . b ) @xmath42 and @xmath39 as functions of @xmath3 calculated for three different values of @xmath9 ( @xmath43).,title="fig:",width=283 ] and the total exciton density @xmath39 as functions of local hybridization @xmath40 calculated by dmrg method for three different values of the interband coulomb interaction @xmath9 ( @xmath41 ) . b ) @xmath42 and @xmath39 as functions of @xmath3 calculated for three different values of @xmath9 ( @xmath43).,title="fig:",width=283 ] fig . 1a shows the bahaviour of the density of zero momentum excitons @xmath42 as a function of local hybridization @xmath40 for several values of the interband coulomb interaction @xmath9 . one can see that both , the local interband hybridization @xmath40 as well as the local coulomb interaction @xmath9 strongly support the condensation of preformed excitons to the zero - momentum state . in all examined cases the density of zero momentum excitons @xmath42 is a monotonically increasing function of @xmath40 which character changes obviously at some critical value of @xmath44 ( @xmath45 , with @xmath46 for @xmath47 and @xmath45 , with @xmath48 for @xmath49 ) . comparing these results with ones obtained in our previous paper for the phase boundary between the cdw and homogeneous phase @xcite , the origin of this different behaviour is obvious . while below @xmath50 the excitonic phase coexists with the @xmath51 phase , above @xmath50 it coexists with the homogeneous one , what leads to a different power law behaviour of @xmath42 for @xmath47 and @xmath49 . unlike @xmath42 , the total exciton density @xmath39 depends only very weakly on @xmath40 ( see the inset in fig . 1a ) indicating that the formation of excitons is primarily driven by the interband coulomb interaction @xmath9 . with regard to the situation in real materials , where always exists a finite overlap of @xmath0 orbitals on neighbouring sites , it is more interesting to ask what happens in the case when a finite @xmath0 bandwidth will be considered . in accordance with some previous theoretical studies , which documented strong effects of the parity of @xmath0 band on the stability of the excitonic phase @xcite we have examined the model for both , the positive ( the even parity ) and negative ( the odd parity ) values of the @xmath0-electron hopping integrals @xmath3 . the results of our non - zero @xmath3 dmrg calculations for @xmath42 are displayed in fig . 1b and they clearly demonstrate that the zero - momentum condensate is suppressed in the limit of positive values of @xmath3 , while it remains robust for negative values of @xmath3 . this result is intuitively expected since our previous hartree - fock ( hf ) results @xcite showed that only the negative values of @xmath3 stabilize the ferroelectric phase , while the positive ones stabilize the antiferroelectric phase . the effect of @xmath3 is especially strong for @xmath9 small , where continuous but very steep changes of @xmath42 are observed for @xmath52 . contrary to this , the total exciton density @xmath39 exhibits only a weak dependence on the @xmath0-electron hopping parameter @xmath3 , over the whole interval of @xmath3 values . till now we have presented results exclusively for @xmath53 . let as now discuss briefly the effect of change of the @xmath0-level position . this study is interesting also from this point of view that taking into account the parametrization between the external pressure and the position of the @xmath0 level ( @xmath54 ) , one can deduce from the @xmath15 dependences of the ground state characteristics also their @xmath55 dependences , at least qualitatively @xcite . the resultant @xmath15 dependences of the density of zero momentum excitons @xmath42 obtained by dmrg method are shown in fig . 2a for several values of @xmath40 and @xmath56 . as a function of @xmath15 calculated by dmrg method for three different values of @xmath40 ( @xmath57 ) . the inset shows the density of @xmath7 electrons @xmath58 near @xmath59 . b ) @xmath60 and @xmath61 as functions of @xmath15 calculated for @xmath62 and @xmath63 . the inset shows the behaviour of @xmath42 and @xmath39 near @xmath64.,title="fig:",width=283 ] as a function of @xmath15 calculated by dmrg method for three different values of @xmath40 ( @xmath57 ) . the inset shows the density of @xmath7 electrons @xmath58 near @xmath59 . b ) @xmath60 and @xmath61 as functions of @xmath15 calculated for @xmath62 and @xmath63 . the inset shows the behaviour of @xmath42 and @xmath39 near @xmath64.,title="fig:",width=283 ] one can see that the density of zero momentum excitons is nonzero over the whole interval of @xmath15 values . moreover , we have found that the values of @xmath42 are extremely enhanced in the region near @xmath65 , what is obviously due to the significant enhancement of the @xmath7 electron population in the @xmath7 band ( see the inset in fig . 2a ) . to describe , in more detail , the process of formation of excitonic bound states with increasing @xmath15 , we have plotted in fig . 2b , besides the density od zero momentum excitons @xmath42 , also the total exciton density @xmath39 , the total @xmath7-electron density @xmath58 and the total density of unbond @xmath7 electrons @xmath61 . it is seen ( see the inset in fig . 2b ) that below @xmath66 , @xmath42 and @xmath39 coincides what means that the excitonic insulator in this region is practically completely driven by the condensation of zero - momentum excitons . above this value @xmath39 starts to increase sharply , while @xmath42 tends to its maximum at @xmath67 and then gradualy decreases to its minimum at @xmath53 . similar behaviour with increasing @xmath15 exhibits also the density of unbond @xmath7 electrons @xmath68 , however the values of @xmath68 are several times larger than @xmath42 . it is interesting to note that although the total exciton density @xmath39 inreases over the whole interval of @xmath15 values , the number of unbond @xmath7 electrns remains practically uchanged over the wide range of @xmath15 values ( from @xmath69 to @xmath70 ) , since its decrease , due to the formation of excitonic pairs , is compensated by the increase of @xmath71 . thus we can conclude that in the pressure induced case , when the @xmath0-level energy shifts up with applied pressure @xcite , the model is able to describe , at least qualitatively , the increase in the total density of excitons with external pressure and the increase or decrease ( according to the initial position of @xmath15 at ambient pressure ) in the @xmath42 and @xmath68 . as already mentioned , physically the most interesting case corresponds however to @xmath72 . the importance of this term emphasizes the fact that the on - site hybridization @xmath40 is actually forbidden in real @xmath7-@xmath0 systems for parity reasons . istead the on - site hybridization , one has to consider in these materials the non - local hybridization with inversion symmetry @xmath73 that leads to @xmath74-dependent hybridization of the opposite parity than corresponds to the @xmath7 band ( @xmath75 ) @xcite . a straightforward extension of the one - dimensional results to two dimensions yields @xmath76 $ ] , where any site on the lattice is given by @xmath77 and @xmath78 is the lattice constant . typical examples of @xmath79 dependence of the excitonic momentum distribution @xmath80 obtained for three representative values of the interband coulomb interaction and two values of @xmath0-electron hopping in the one dimension are displayed in fig . 3a and fig . 3b . as a function of @xmath79 calculated by dmrg method for three different values of @xmath9 and two different values of @xmath3 ( @xmath81 ) . , title="fig:",width=283 ] as a function of @xmath79 calculated by dmrg method for three different values of @xmath9 and two different values of @xmath3 ( @xmath81 ) . , title="fig:",width=283 ] these results clearly demonstrate that there is no sign of divergence in the @xmath79-dependence of @xmath82 neither for @xmath83 nor for @xmath84 and thus there is no signal of forming the bose - einstein condensate in the presence of non - local hybridization with the inversion symmetry . on the base of these results we can conclude that possible candidates for the appearance of the bose - einstein condensation of excitons in real @xmath7-@xmath0 materials , are only systems with local hybridization that supports the formation of the bose - einstein condensate , but not systems with non - local hybridization ( of the inversion symmetry ) , what strongly limits the class of materials , where this phenomenon could be observed . on the other hand , it should be noted that although the local hybridization is forbidden in real @xmath7-@xmath0 electron systems , there is still a possibility to induce a finite local hybridization by some additional , not fully electronic mechanisms . such an additional mechanism could be , for example , the electron - phonon interaction @xmath85 . as shown by menezes at al . @xcite such an interaction can be reduced to the phonon - mediated local hybridization that can also lead to the formation of bose - einstein condensate under the supposition that it will be sufficiently strong , what depends on values of electron - phonon constants . since not all of the above mentioned candidates are the one - dimensional or quasi one - dimensional systems , it is interesting to ask if the above obtained results persist also in higher dimensions . unfortunately , the dmrg method is not a very convenient tool for the study of lattice systems in @xmath32 . therefore , it is necessary to choose another method that should be able to describe sufficiently precise the excitonic effects in @xmath7-@xmath0 systems . from this point of view a very promising method seems to be the hf approximation with the cdw instability . indeed , we have showed in our previous paper @xcite , that this method is able to describe almost perfectly the ground - state phase diagram of the extended falicov - kimball model in @xmath86 , including the existence of excitonic phase . let us briefly summarize the basic steps of this approximation . in the presence of cdw instability , the order parameters can be written as follows @xcite : @xmath87 where @xmath88 and @xmath89 are the order parameters of the cdw state for the @xmath7- and @xmath0-electrons , @xmath90 is the excitonic average and @xmath91 ( @xmath92 ) is the nesting vector for @xmath93 ( @xmath86 ) . with @xmath98 and the corresponding dispersions @xmath99 , @xmath100 and @xmath101 are obtained by the fourier transform of the @xmath7/@xmath0-electron hopping amplitudes and the local / nonlocal hybridization . then the hf parameters @xmath102 and the density of zero - momentum excitons @xmath42 can be expressed directly in terms of the bogoliubov - de gennes eigenvectors @xmath103 @xcite . to verify the ability of this method to describe the formation of the bose - einstein condensate we have calculated firstly the density of the zero momentum excitons as functions of model parameters @xmath104 and @xmath15 for the one dimensional case . the results of our numerical calculations are displayed in fig 4 ( the left panels ) and they show that the hf approximation with the cdw instability is able to decribe all aspects of formation of zero - momentum condensate in @xmath93 discussed above within the dmrg method . although , there are some small qualitative as well as quantitative differences , like the discontinuous changes of @xmath42 with @xmath3 near @xmath83 , or almost two times larger values of @xmath50 in comparison to dmrg studies ( see the inset in the a ) panel ) , in principle , the dmrg and hf pictures are same . in accordance with dmrg results we have found a strong enhancement of @xmath42 near @xmath105 also in the hf solutions . performing the numerical derivative of @xmath58 with respect to @xmath15 and comparing it with the behaviour of @xmath106 one can easily verify ( see the instet in the e ) panel ) that this enhancement is indeed connected with changes in the occupation of @xmath0 ( @xmath7 ) orbitals as conjectured above on the base of @xmath107 results . as a function of @xmath108 calculated by hf approximation with the @xmath51 instability in the one ( the left panels ) and two dimensions ( the right panels ) . the insets in the panels a ) and b ) show the phase boundary between the homogeneous excitonic insulator ( hei ) phase and the coexisting excitonic insulator and cdw state ( cdw+ei).,title="fig:",width=283 ] as a function of @xmath108 calculated by hf approximation with the @xmath51 instability in the one ( the left panels ) and two dimensions ( the right panels ) . the insets in the panels a ) and b ) show the phase boundary between the homogeneous excitonic insulator ( hei ) phase and the coexisting excitonic insulator and cdw state ( cdw+ei).,title="fig:",width=283 ] as a function of @xmath108 calculated by hf approximation with the @xmath51 instability in the one ( the left panels ) and two dimensions ( the right panels ) . the insets in the panels a ) and b ) show the phase boundary between the homogeneous excitonic insulator ( hei ) phase and the coexisting excitonic insulator and cdw state ( cdw+ei).,title="fig:",width=283 ] as a function of @xmath108 calculated by hf approximation with the @xmath51 instability in the one ( the left panels ) and two dimensions ( the right panels ) . the insets in the panels a ) and b ) show the phase boundary between the homogeneous excitonic insulator ( hei ) phase and the coexisting excitonic insulator and cdw state ( cdw+ei).,title="fig:",width=283 ] as a function of @xmath108 calculated by hf approximation with the @xmath51 instability in the one ( the left panels ) and two dimensions ( the right panels ) . the insets in the panels a ) and b ) show the phase boundary between the homogeneous excitonic insulator ( hei ) phase and the coexisting excitonic insulator and cdw state ( cdw+ei).,title="fig:",width=283 ] as a function of @xmath108 calculated by hf approximation with the @xmath51 instability in the one ( the left panels ) and two dimensions ( the right panels ) . the insets in the panels a ) and b ) show the phase boundary between the homogeneous excitonic insulator ( hei ) phase and the coexisting excitonic insulator and cdw state ( cdw+ei).,title="fig:",width=283 ] the same calculations we have performed also in d=2 . the resultant behaviours of the density of zero momentum excitons @xmath42 as functions of @xmath104 and @xmath15 are plotted in fig . 4 ( the right panels ) and they clearly confirm that with increasing dimension of the system the zero - momentum condensate remains robust . in comparison to the one - dimensional results there are two differences , and namely , that the discontinuous changes of @xmath42 ( as a function of @xmath3 ) take place now strictly at @xmath83 in all examined cases and that the maximum in @xmath106 shifts to higher values of @xmath15 , where the periodic solutions with cdw instability minimize the ground state energy . and finally , it should be noted that in accordance with dmrg results we have found no sign of formation of bose - einstein condensate for the case of nonlocal hybridization with the inversion symmetry neither in our hf solutions for both @xmath93 as well as @xmath86 . in conclusion , we have examined effects of various factors , like @xmath0-electron hopping , the local and nonlocal hybridization , as well as the increasing dimension of the system on the formation and condensation of excitonic bound states in the generalized falicov - kimball model . it was found that the negative values of the @xmath0-electron hopping integrals @xmath3 support the formation of zero - momentum condensate , while the positive values of @xmath3 have the fully opposite effect . the opposite effects on the formation of condensate exhibit also the local and nonlocal ( with the inversion symmetry ) hybridization . the first one strongly supports the formation of zero - momentum condensate , while the second one destroys it completely . moreover , it was found that in the pressure induced case ( @xmath109 ) , the model is able to describe , at least qualitatively , the increase in the total density of excitons @xmath39 with external pressure and the increase or deccrease ( according to the initial position of @xmath15 at ambient pressure ) in @xmath42 and @xmath68 . and finaly , it was shown ( by hf studies ) that the zero - momentum condensate remains robust with increasing dimension of the system . c. monney , g. monney , p. aebi , and h. beck , phys . rev . b*85 * , 235150 ( 2012 ) ; b. zenker , h. fehske , h. beck , c. monney , and a.r . bishop , phys . rev . b*88 * , 075138 ( 2013 ) ; g. monney , c. monney , b. hildebrand , p. aebi , and h. beck , phys . lett . * 114 * , 086402 ( 2015 ) ; h. watanabe , k. seki , and s. yunoki , phys . * , 205135 ( 2016 ) .

the density - matrix - renormalization - group ( dmrg ) method and the hartree - fock ( hf ) approximation with the charge - density - wave ( cdw ) instability are used to study a formation and condensation of excitonic bound states in the generalized falicov - kimball model . in particular , we examine effects of various factors , like the @xmath0-electron hopping , the local and nonlocal hybridization , as well as the increasing dimension of the system on the excitonic momentum distribution @xmath1 and especially on the number of zero momentum excitons @xmath2 in the condensate . it is found that the negative values of the @xmath0-electron hopping integrals @xmath3 support the formation of zero - momentum condensate , while the positive values of @xmath3 have the fully opposite effect . the opposite effects on the formation of condensate exhibit also the local and nonlocal hybridization . the first one strongly supports the formation of condensate , while the second one destroys it completely . moreover , it was shown that the zero - momentum condensate remains robust with increasing dimension of the system .

the prompt localization of gamma - ray bursts ( grbs ) by bepposax led to the discovery of x - ray / optical / radio afterglows and associated host galaxies . subsequent detections of absorption and emission features at high redshifts ( @xmath1 ) in optical afterglows of grb and their host galaxies clearly demonstrate that at least some of the grb sources lie at cosmological distances ( for reviews , see @xcite ; @xcite ) . a common feature of all acceptable models of cosmological @xmath2-ray bursters is that a relativistic wind is a source of grb radiation . the lorentz factor , @xmath3 , of such a wind is about @xmath4 or even more ( e.g. , @xcite ; @xcite ) . a very strong magnetic field may be in the plasma outflowing from cosmological @xmath2-ray bursters ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . a relativistic , strongly magnetized wind interacts with an ambient medium ( e.g. an ordinary interstellar gas ) and decelerates . it was pointed out ( @xcite ) that such an interaction , assumed to be shock - like , may be responsible for generation of cosmological grbs . the interaction between a relativistic , strongly magnetized wind and an ambient plasma was studied numerically by smolsky & usov ( 1996 , 2000 ) and usov & smolsky ( 1998 ) . these studies showed that about 70% of the wind energy is transferred to the ambient plasma protons that are reflected from the wind front . the other @xmath5% of the wind energy losses is distributed between high - energy electrons and low - frequency electromagnetic waves that are generated at the wind front because of nonstationarity of the wind ambient plasma interaction ( see below ) . high - energy electrons , accelerated at the wind front and injected into the region ahead of the front , generate synchro - compton radiation in the fields of the low - frequency waves . this radiation closely resembles synchrotron radiation and can reproduce the non - thermal radiation of grbs observed in the ginga and batse ranges ( from a few kev to a few mev ) . @xcite and @xcite suggested that grb might be sources of radio emission , and that it might be used to determine their distances and , through the dispersion measure , the density of the intergalactic plasma . here we consider some properties of the low - frequency waves generated at the wind front . we argue that coherent emission by the high - frequency tail of these waves may be detected , in addition to the high - frequency ( x - ray and @xmath2-ray ) emission of grbs , as a short pulse of low - frequency radio emission ( @xcite ) . our mechanism for production of short pulses of low - frequency radio emission from relativistic , strongly magnetized wind - generated cosmological grbs applies very generally . for the sake of concreteness , we consider wind parameters that are natural in a grb model that involves a fast rotating compact object like a millisecond pulsar or dense transient accretion disc with a surface magnetic field @xmath6 g ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . in this model the rotational energy of compact objects is the energy source of cosmological grbs . the electromagnetic torque transfers this energy on a time scale of seconds to the energy of a poynting flux - dominated wind that flows away from the object at relativistic speeds , @xmath7 ( e.g. , @xcite ) . the wind structure at a time @xmath8 is similar to a shell with radius @xmath9 and thickness of @xmath10 , where @xmath11 is the characteristic deceleration time of the compact object s rotation , or the dissipation time of a transient accretion disc . the strength of the magnetic field at the front of the wind may be as high as @xmath12 where @xmath13 cm is the radius of the compact object , @xmath14 is the angular velocity at the moment of its formation and @xmath15s@xmath16 cm is the radius of the light cylinder . the distance at which deceleration of the wind due to its interaction with an ambient gas becomes important is @xmath17 where @xmath18 is the density of the ambient gas and @xmath19 is the initial kinetic energy of the outflowing wind . ( 2 ) assumes spherical symmetry ; for beamed flows @xmath19 is @xmath20 times the wind energy per steradian . at @xmath21 , the main part of @xmath22 is lost by the wind in the process of its inelastic interaction with the ambient medium , and the grb radiation is generated . substituting @xmath23 for @xmath24 into equation ( [ bobj ] ) , we have the following estimate for the magnetic field at the wind front at @xmath25 : @xmath26 where we have introduced a parameter @xmath27 which gives the fraction of the wind power remaining in the magnetic field at the deceleration radius . for plausible parameters of cosmological @xmath2-ray bursters , @xmath28 g , @xmath14 s@xmath29 , @xmath30 ergs , @xmath7 and @xmath31 @xmath32 , from equation ( [ bdec ] ) we have @xmath33 g. for consideration of the interaction between a relativistic magnetized wind and an ambient gas , it is convenient to switch to the co - moving frame of the outflowing plasma ( the wind frame ) . while changing the frame , the magnetic and electric fields in the wind are reduced from @xmath34 and @xmath35^{1/2 } \simeq b$ ] in the frame of the @xmath2-ray burster to @xmath36 and @xmath37 in the wind frame . this is analogous to the the well - known transformation of the coulomb field of a point charge : purely electrostatic in the frame of the charge , but with @xmath38 in a frame in which the charge moves relativistically . using this and equation ( [ bdec ] ) , for typical parameters of cosmological @xmath2-ray bursters we have @xmath39 g at @xmath40 . in the wind frame , the ambient gas moves to the wind front with the lorentz factor @xmath3 and interacts with it . the main parameter which describes the wind ambient gas interaction is the ratio of the energy densities of the ambient gas and the magnetic field , @xmath41 , of the wind @xmath42 where @xmath43 is the density of the ambient gas in the wind frame and @xmath44 is the proton mass . at the initial stage of the wind outflow , @xmath45 , @xmath46 , but it increases in the process of the wind expansion as @xmath41 decreases . when @xmath47 is @xmath48 ( at @xmath49 , the interaction between the wind and the ambient gas is strongly nonstationary , and effective acceleration of electrons and generation of low - frequency waves at the wind front both begin ( @xcite , 2000 ; @xcite ) . for @xmath50 , the mean lorentz factor of outflowing high - energy electrons @xmath51 accelerated at the wind front and the mean field of low - frequency waves @xmath52 weakly depend on @xmath47 ( see table 1 ) and are approximately given by @xmath53 to within a factor of 2 , where @xmath54 and @xmath55 are the magnetic and electric field components of the waves . the mechanism of generation of these waves is coherent and consists of the following : at the wind front there is a surface current that separates the wind matter with a very strong magnetic field and the ambient gas where the magnetic field strength is negligible . this current varies in time because of nonstationarity of the wind - ambient gas interaction and generates low - frequency waves . at @xmath56 , for the bulk of high - energy electrons in the region ahead of the wind front the characteristic time of their synchrotron energy losses is much less than the grb duration . in this case , the luminosity per unit area of the wind front in @xmath2-rays is @xmath57 while the same luminosity in low - frequency waves is @xmath58 . using these , the ratio of the luminosity in low - frequency waves and the @xmath2-ray luminosity is @xmath59 from table 1 , we can see that the grb light curves in both @xmath2-rays and low - frequency waves have maximum when @xmath47 is about 0.4 , and the maximum flux in low - frequency waves is about two times smaller than the maximum flux in @xmath2-rays ( @xmath60 ) . at @xmath61 the value of @xmath62 decreases with increasing @xmath47 . therefore , we expect that the undispersed duration @xmath63 of the low - frequency pulse to be somewhat smaller than the grb duration , and the energy fluence in low - frequency waves to be roughly an order of magnitude smaller than the energy fluence in @xmath2-rays . the rise time of the radio pulse is very short because that the value of @xmath64 increases very fast when @xmath47 changes from 0.3 to 0.4 ( see table i ) . figure 1 shows a typical spectrum of low - frequency waves generated at the wind front in the wind frame . this spectrum has a maximum at the frequency @xmath65 which is about three times higher than the proton gyrofrequency @xmath66 in the wind field @xmath41 . taking into account the doppler effect , in the observer s frame the spectral maximum for low - frequency waves is expected to be at the frequency @xmath67 where @xmath68 is the cosmological redshift . for typical parameters of cosmological grbs , @xmath39 g and @xmath69 , we have @xmath70 mhz . unfortunately , the bulk of the low - frequency waves is at low frequencies , and can not be observed . however , their high - frequency tail may be detected . at high frequencies , @xmath71 , the spectrum of low - frequency waves may be fitted by a power law ( @xcite ) : @xmath72 where @xmath73 . in the simulations of the wind ambient gas interaction ( @xcite , 2000 ; @xcite ) both the total numbers of particles of the ambient gas and the sizes of spatial grid cells are restricted by computational reasons , so that the the spectrum ( [ bnu ] ) is measured reliably only at @xmath74 . the amplitudes of the computed oscillations with @xmath75 are so small ( @xmath76 ) that they can not be distinguished from computational noise ( @xcite ) . future calculations with greater computational resources may alleviate this problem . the value of @xmath77 depends on many parameters of both the grb bursters and the ambient gas around them , and its estimate , @xmath70 mhz , is uncertain within a factor of 23 or so . in the most extreme case in which @xmath77 is as high as a few mhz , the high - frequency tail of low - frequency waves may be continued up to @xmath5 mhz where ground - based radio observations may be performed . in this case , the energy fluence in a pulse of radio emission at @xmath78 mhz may be as high as a few percent of the grb energy fluence in @xmath2-rays . a pulse of low - frequency radio emission is strongly affected by intergalactic plasma dispersion in the process of its propagation . at the frequency @xmath79 , the radio pulse retardation time with respect to a grb is @xmath80 where @xmath81 is the intergalactic dispersion measure in electrons/@xmath82 , @xmath83 is the group velocity of radio emission , @xmath84 is the refractive index and @xmath79 in hz . from equation ( [ taunu ] ) , for the plausible parameters of @xmath85 @xmath32 and a distance of @xmath86 cm , at @xmath87 mhz we have @xmath88 s. this is time enough to steer a radio telescope for the radio pulse detection . in equation ( [ taunu ] ) , we neglected the radio pulse retardation time in our galaxy , which is typically one or two orders of magnitude less than that in the intergalactic gas . the observed duration of the low - frequency pulse at the frequency @xmath79 in the bandwidth @xmath89 is @xmath90\ , . \label{deltatnu}\ ] ] for plausible parameters , @xmath91 mhz , @xmath92 mhz , @xmath93 s and @xmath94 s , we have @xmath95 s ; the observed duration of low - frequency radio pulses is determined by intergalactic plasma dispersion , except for extremely long grbs . it is now possible to estimate , given assumed values for the magnetic field , the amplitude of the signal produced . we also assume that the plasma field couples efficiently to the free space radiation field . for a radio fluence @xmath96 and a radio fluence spectral density @xmath97 the radio spectral flux density is @xmath98 for the latter ( dispersion - limited ) case with the parameters @xmath99 erg @xmath100 , @xmath101 , @xmath102 , @xmath103 mhz , @xmath87 mhz , @xmath104 mhz , @xmath105 s we find @xmath106 jy . the appropriate value of @xmath107 is very uncertain . in some models it may be @xmath108 , but for grb with sharp subpulses its value is limited by the requirement that the magnetic stresses not disrupt the thinness of the colliding shells ( @xcite ) . for subpulses of width @xmath109 of the grb width this suggests @xmath110 ; typical estimates are @xmath111 and @xmath112 , leading to @xmath113 jy . these large values of @xmath114 may be readily detectable , although the assumed values of @xmath107 are very uncertain . there are additional uncertainties . we have assumed that the radio pulse spectrum ( 11 ) is valid up to a frequency @xmath115 mhz that may be hundreds of times higher than @xmath77 . as discussed above , the spectrum ( 11 ) of low - frequency waves is calculated directly only at @xmath116 . at @xmath117 , the spectrum must be extrapolated , with unknown confidence , from the calculations . the radio spectral flux density at @xmath115 mhz may therefore be less than the preceding estimates . however , even in this case the very high sensitivity of measurements at radio frequencies may permit the detection of coherent low - frequency radio emission from grb . in this letter , we have shown that grbs may be accompanied by very powerful short pulses of low - frequency radio emission . for detection of these radio pulses it may be necessary to perform observations at lower frequencies than are generally used in radio astronomy , which are limited by the problem of transmission through and refraction by the ionosphere . in particular , observations from space are free of ionospheric refraction and are shielded by the ionosphere from terrestrial interference . although even harder to predict , detection from the ground at higher frequencies may also be possible . space observations are possible at frequencies down to that at which the interstellar medium becomes optically thick to free - free absorption . this frequency is @xmath118 where @xmath119 @xmath32 ) and @xmath120k ( @xcite ) , @xmath121 and @xmath122 are the interstellar electron density and temperature , respectively , and @xmath123 is the galactic latitude . the expression @xmath124 may be @xmath108 , but could be substantially larger if the electrons are strongly clumped or cold . on the other hand , although much of the interstellar volume is filled with very hot ( @xmath125k ) and radio - transparent gas , this probably contains very little of the electron column density . intergalactic absorption poses a somewhat less restrictive condition if the medium is hot ( @xmath125k ) , as generally assumed . at these low frequencies , and even at tens of mhz , interstellar scintillation ( @xcite ) will be very large . observations of coherent radio emission from grb would not only illuminate the physical conditions in their radiating regions , but would determine ( through the dispersion measure ) the mean intergalactic plasma density and ( through the scintillation ) its spatial structure . the flux density implied by eq . ( [ rflux ] ) appears impressively large , but it applies only to the brief period when the dispersed signal is sweeping through the bandwidth @xmath126 of observation , so that it is unclear if it is , in fact , excluded by the very limited data ( @xcite ) available . further , the extrapolation of the radiated spectrum to @xmath127 is very uncertain . finally , we have also not considered the ( difficult to estimate ) temporal broadening of this brief transient signal by intergalactic scintillation , which will both reduce its amplitude and broaden its time - dependence . searches for radio pulses started about 50 years ago , prior to the discovery of grbs . during wide beam studies of ionospheric scintillations , simultaneous bursts at 45 mhz of @xmath128 s duration were reported by smith ( 1950 ) at sites 160 km apart . these events were detected at night , approximately once a week . the origin of these bursts was never determined . the results of smith ( 1950 ) were not confirmed by subsequent observations at frequencies of 150 mhz or higher . modern observations ( @xcite at 151 mhz ; @xcite at 74 mhz ; @xcite broadband ; see @xcite for a review ) set upper bounds to the brightnesses of some grb at comparatively high frequencies . these bounds are not stringent , and do not exclude extrapolations of the lower frequency fluxes suggested here . there are few data at lower frequencies . power spectrum of low - frequency electromagnetic waves generated at the front of the wind in the wind frame in a simulation with @xmath130 g , @xmath131 , and @xmath132 . the spectrum is fitted by a power law ( dashed line ) .

gamma - ray bursts , if they are generated in the process of interaction between relativistic strongly magnetized winds and an ambient medium , may be accompanied by very short pulses of low - frequency radio emission . the bulk of this emission is expected to be at the frequencies of @xmath0 mhz and can not be observed . however , the high - frequency tail of the low - frequency radio emission may reach a few ten mhz and be detected , especially if the strength of the magnetic field of the wind is extremely high .

let @xmath1 denote the grassmannian of @xmath2-planes in @xmath3 . it has a stratification by _ positroid varieties _ @xmath4 @xcite , where @xmath5 ranges over the finite set @xmath6 of @xmath7-bounded affine permutations ( defined in section [ sec : main ] ) . each positroid variety is the intersection of @xmath8 cyclically rotated schubert varieties . in @xcite , knutson , lam , and speyer identified the cohomology class of a positroid variety with the _ affine stanley symmetric function _ @xmath9 @xcite . the totally nonnegative part @xmath10 of the real grassmannian is the locus where all plcker coordinates take nonnegative values @xcite , and was studied extensively by postnikov . arkani - hamed and trnka @xcite , motivated by the study of scattering amplitudes in @xmath0 super yang - mills , proposed that @xmath10 should be considered a grassmannian - analogue of a simplex . arbitrary convex polytopes are images of simplices under affine or linear maps , and arkani - hamed and trnka proposed to study the _ amplituhedron _ : the image of the totally nonnegative grassmannian induced by a linear map @xmath11 ( which in turn gives a rational map @xmath12 ) . in addition , physical considerations suggested the study of triangulations of the amplituhedron , obtained as unions of images of the positroid cells @xmath13 , again under the map @xmath14 . specifically , the _ scattering amplitude _ can be obtained by summing differential forms over cells of a triangulation of the amplituhedron . the behavior of positroid cells under the map @xmath14 exhibit a number of features not present in usual convex geometry , including : 1 . even when @xmath15 is generic , the image @xmath16 may not have the expected dimension . for example , even if @xmath17 we may have @xmath18 for generic @xmath15 . the map @xmath19 can be dimension - preserving , but have degree @xmath20 greater than one . in this paper , we study the complex geometry of the behavior of the stratification @xmath21 under the map @xmath14 , from a schubert calculus perspective . let @xmath22 denote the closure of the image of @xmath4 under @xmath14 . we call @xmath22 an _ amplituhedron variety _ in the case that it has the same dimension as @xmath4 . recall that the cohomology - coefficients . ] ring @xmath23 can be identified with a quotient of the ring of symmetric functions , and that the basis of schubert classes correspond to the schur functions @xmath24 , labeled by partitions @xmath25 that fit inside a @xmath26 rectangle . let @xmath27 . for @xmath28 we let @xmath29 be the partition obtained from @xmath30 by adding @xmath31 columns of height @xmath2 to the left of @xmath30 . for example , with @xmath32 and @xmath33 , we may have @xmath34{&&&\\&&\\ & \bl \\ & \bl } } } \qquad \qquad \qquad \mu^+ = { { \tableau[py]{&&&&&\\&&&&\\&&\\&&}}}\ ] ] given @xmath35 representing a cohomology class in @xmath36 , we define the _ truncation _ @xmath37 by @xmath38 let @xmath39 denote the degree of the map @xmath40 . the cohomology class of the amplituhedron variety @xmath22 is equal to @xmath41 . we also prove that @xmath42 if and only if @xmath43 . as a corollary , we deduce a criterion for @xmath16 to have the same dimension as @xmath44 , which corresponds to the physical notion of kinematical support " . we also obtain some estimates on the degree @xmath39 . we present a number of possible further directions in section [ sec : other ] . * * my understanding of the amplituhedron owes a lot to nima arkani - hamed and jara trnka , who have taught me much through many conversations . i thank jake bourjaily for discussions related to scattering amplitudes , and for his `` positroids '' package in mathematica . i benefited from discussions with allen knutson and alex postnikov , and am especially grateful to david speyer for a number of helpful comments . much of what i understand about scattering amplitudes was learnt during a reading seminar with henriette elvang , yu - tin huang , cindy keeler , tim olson , sam roland , and david speyer . i thank them all for teaching me this subject . we fix positive integers @xmath45 satisfying @xmath46 , and we let @xmath27 . let @xmath47 denote the space of @xmath48 matrices , and let @xmath49 denote the open subset of full rank @xmath48 matrices . we think of @xmath50 as a linear map @xmath51 . the map @xmath15 induces a rational map @xmath52 given by @xmath53 . the _ exceptional locus _ @xmath54 of @xmath14 is the subset of @xmath1 where the map @xmath14 is not defined : @xmath55 here @xmath56 is the usual kernel of a linear map . the exceptional locus @xmath54 is in fact a schubert variety which has codimension @xmath57 . [ lem : bundle ] the morphism @xmath58 is a fiber bundle with fiber @xmath59 . we use the @xmath60 actions on @xmath15 and on @xmath1 to reduce to the case that @xmath15 is the orthogonal projection of @xmath61 onto @xmath62 . then the map @xmath14 looks like @xmath63 where @xmath64 is a @xmath65 matrix representing a point in @xmath66 , and @xmath67 denotes the @xmath59 fiber . we use the notation @xmath68:=\{1,2,\ldots , a\}$ ] . let @xmath69}{k}$ ] be a @xmath2-element subset of @xmath70 $ ] . let @xmath71 be a flag in @xmath3 , so that @xmath72 . the schubert variety @xmath73 is given by @xmath74 ) \text { for all } j \in [ n]\}.\ ] ] thus @xmath75}(f_\bullet ) = \gr(k , n)$ ] and @xmath76 , where @xmath77 . here and elsewhere , we always mean complex ( co)dimension when referring to complex subvarieties . let @xmath78 be a flag in @xmath79 . then @xmath80 is the partial flag @xmath81 of subspaces with successive dimensions @xmath82 . we denote by @xmath83 a schubert variety in @xmath66 , where @xmath84 is a @xmath2-element subset of @xmath85 $ ] . a full - flag extension of @xmath80 is simply any flag @xmath86 in @xmath3 whose @xmath82-dimensional pieces give @xmath80 . [ lem : inverse ] we have @xmath87 where @xmath86 is any full - flag extension of @xmath80 , and @xmath88 $ ] is considered a subset of @xmath70 $ ] via the natural inclusion @xmath85 = \{1,2,\ldots , k+m\ } \subset \{1,2,\ldots , n\}=[n]$ ] . suppose @xmath89 . then @xmath90)$ ] , and so @xmath91)$ ] for all @xmath92 $ ] . that is , @xmath93)$ ] for all @xmath94 $ ] . since membership in @xmath73 imposes no condition on @xmath95 for @xmath96 $ ] , we conclude that @xmath97 . but using lemma [ lem : bundle ] , we see that @xmath98 and @xmath73 are closed irreducible subvarieties of @xmath1 of the same dimension , and so must be identical . if @xmath99 then @xmath83 is a single point @xmath100 . lemma [ lem : inverse ] then says that @xmath101 is a subgrassmannian of @xmath1 . we shall need the following version of kleiman transversality . [ thm : kleiman ] assume the base field is @xmath102 . let @xmath103 be an integral algebraic scheme with a transitive action of an algebraic group @xmath104 . let @xmath105 be integral subschemes . then 1 . there exists a dense subset @xmath106 such that for @xmath107 , the intersection @xmath108 is proper , that is , each component has dimension @xmath109 . 2 . if in addition @xmath64 and @xmath15 are smooth , then @xmath110 can be chosen so that for all @xmath107 , the subschemes @xmath111 and @xmath15 intersect transversally , that is , the intersection @xmath108 is smooth and each component has dimension @xmath109 . we remark that if @xmath111 and @xmath15 intersect transverally then the intersection @xmath108 , being smooth , must be contained in the smooth locus of both @xmath111 and @xmath15 . we shall also need the following technical result which appears in the proof of theorem [ thm : kleiman ] . [ lem : kleiman ] assume the base field is @xmath102 . consider a diagram with integral algebraic schemes : \(m ) [ matrix of math nodes , row sep=3em , column sep=2.5em , text height=1.5ex , text depth=0.25ex ] & w & & z + s & & x & + ; ( m-1 - 2 ) edge node[description ] @xmath112 ( m-2 - 1 ) ; ( m-1 - 2 ) edge node[description ] @xmath113 ( m-2 - 3 ) ; ( m-1 - 4 ) edge node[description ] @xmath114 ( m-2 - 3 ) ; 1 . assume @xmath113 is flat . then , there exists a dense open subset @xmath110 of @xmath115 such that for each point @xmath116 , either the fibered product , @xmath117 , is empty or it is equidimensional and its dimension is given by the formula , @xmath118 2 . assume @xmath113 is flat with smooth fibers . assume @xmath15 is smooth . then @xmath117 is smooth for each point @xmath119 in an open dense subset of @xmath115 . let @xmath120 be an irreducible subvariety . for @xmath50 , we define @xmath121 for a generic @xmath15 , the subscheme @xmath122 is irreducible and dense in @xmath123 . thus @xmath124 is itself an irreducible subvariety . there is a @xmath60-action on @xmath1 and a @xmath60-action on @xmath49 . we choose compatible conventions so that @xmath125 . let @xmath126 be a schubert subvariety . for @xmath50 a full - rank matrix , let @xmath127 be as in lemma [ lem : inverse ] . [ lem : main ] fix @xmath128 an irreducible subvariety , @xmath129 a schubert variety , satisfying @xmath130 . then there exists a zariski - open subset @xmath131 such that : 1 . for all @xmath132 , we have @xmath122 is open and dense in @xmath123 ; 2 . 1 . either for all @xmath132 we have @xmath133 and @xmath134 , 2 . or for all @xmath132 we have @xmath135 , the intersection @xmath136 is transversal , and all intersection points lie in the locus inside @xmath137 where the map @xmath138 has fibers of cardinality exactly @xmath139 , where @xmath139 is the degree of the map @xmath140 . furthermore , @xmath139 is constant for all @xmath132 . 3 . for all @xmath132 , we have @xmath123 intersects @xmath141 transversally , and all intersection points lie in @xmath122 . in the following , we shall use the fact that a morphism between irreducible varieties is generically flat , and a morphism between smooth irreducible varieties is generically smooth . ( here all varieties are over @xmath102 . ) similar results are used throughout @xcite , and we refer the reader there for precise references . for a fixed full - rank @xmath142 , by theorem [ thm : kleiman ] , there exists an open subset @xmath143 such that @xmath123 and @xmath144 intersect properly for any @xmath145 . by dimension considerations we will have @xmath146 is open and dense in @xmath123 for any @xmath145 . but @xmath125 . the map @xmath147 gives a surjective map from @xmath60 to @xmath49 . it follows that the image of @xmath148 contains a zariski open subset @xmath149 . define @xmath150 by @xmath151 obviously @xmath152 is an irreducible and open subset of @xmath153 , and the fiber of @xmath152 over @xmath154 is @xmath122 . define @xmath155 by @xmath156 . define @xmath157 obviously , @xmath115 is an irreducible subset of @xmath158 . we now assume that @xmath159 so that for @xmath15 in a dense open subset of @xmath160 , we have @xmath161 , and we are in case ( b ) of part ( 2 ) of the lemma . ( the case where @xmath162 is easier since one expects @xmath124 and @xmath163 not to intersect . ) let @xmath20 be the degree of the map @xmath164 . for a dense open subset @xmath165 , the fiber @xmath166 for @xmath167 will have exactly @xmath20 points . let @xmath168 denote the closure of @xmath115 in @xmath158 . over an open subset @xmath169 , the map @xmath170 will be flat , and the fiber over @xmath171 will be reduced and equal to @xmath124 , and all these fibers will have the same dimension . by shrinking @xmath172 if necessary , we may assume that for @xmath171 , the fiber @xmath173 is an open dense subset of @xmath124 that is contained in @xmath174 , with the additional property that for @xmath175 , the fiber @xmath176 has cardinality equal to exactly @xmath20 . let @xmath177 . by replacing @xmath178 by its smooth locus , and shrinking @xmath172 if necessary , we may in addition assume that @xmath178 is smooth . for @xmath179 , we have @xmath180 and @xmath181 . thus if the above properties hold for @xmath182 and @xmath183 , they also hold for @xmath184 and @xmath185 ( where @xmath186 sends @xmath187 to @xmath188 ) . so we may assume that @xmath172 and @xmath178 are closed under the @xmath104-action . consider the natural map @xmath189 . since @xmath178 is smooth and irreducible , @xmath113 is flat with smooth fibers over a dense subset @xmath148 of @xmath66 . the map @xmath113 commutes with the actions of @xmath104 on @xmath190 and on @xmath66 . since @xmath104 acts transitively on @xmath66 , the translations @xmath191 for @xmath192 obviously cover @xmath66 , and it follows that @xmath113 is flat over @xmath66 . now we apply lemma [ lem : kleiman](1,2 ) to the family \(m ) [ matrix of math nodes , row sep=3em , column sep=2.5em , text height=1.5ex , text depth=0.25ex ] & t & & y^_i + u_2 & & ( k , k+m ) & + ; ( m-1 - 2 ) edge node[description ] @xmath112 ( m-2 - 1 ) ; ( m-1 - 2 ) edge node[description ] @xmath113 ( m-2 - 3 ) ; ( m-1 - 4 ) edge node[description ] @xmath193 ( m-2 - 3 ) ; where @xmath194 is the inclusion of the smooth locus @xmath195 . we deduce that there is a dense open subset @xmath196 such that for each @xmath197 , we have that @xmath198 and @xmath195 intersect transversally . to finish obtaining the statement of ( 2)(b ) , it remains to show that we can find @xmath199 so that for @xmath200 , we have @xmath201 . that is , there are no intersection points in @xmath202 or @xmath203 . to do so , we repeat the argument ( using lemma [ lem : kleiman](1 ) ) for the family @xmath204 and the inclusion @xmath205 . the typical fiber of @xmath204 has lower dimension than @xmath124 since @xmath198 is open dense in @xmath124 . thus we expect @xmath202 not to intersect @xmath163 . we deduce that there exists a dense open subset of @xmath206 where all the intersection points of @xmath124 and @xmath163 lie in @xmath207 . repeating the argument , we can also find a dense open subset of @xmath206 such that the intersection points of @xmath124 and @xmath163 lie in @xmath208 . thus we can find @xmath209 satisfying conditions ( 1 ) and ( 2 ) of the lemma . finally , the condition ( 3 ) holds in an open subset @xmath210 : the argument here only requires applying theorem [ thm : kleiman](1,2 ) . we then set @xmath211 . an irreducible subvariety @xmath128 of complex codimension @xmath20 has a cohomology class @xmath212 \in h^{2d}(\gr(k , n))$ ] , which must be non - zero . transverse intersections allow one to compute products in cohomology . [ thm : ful ] let @xmath103 be a nonsingular variety . let @xmath105 be closed irreducible subvarieties . suppose @xmath64 and @xmath15 intersect transversally . then we have @xmath213 \cdot [ z ] = [ y \cap z]\ ] ] in the cohomology ring @xmath214 . when @xmath215 is a finite set of @xmath114 ( reduced ) points , we have @xmath216 = r[\pt ] \in h^*(x)$ ] . let @xmath217 be the standard flag in @xmath3 . the cohomology ring @xmath36 vanishes in odd degrees , and the set @xmath218 \mid \codim(x_i ) = d\}$ ] of schubert classes forms a @xmath219-basis of @xmath220 . recall that @xmath36 is isomorphic to the quotient of the ring @xmath221 of symmetric functions by an ideal @xmath222 ( see @xcite ) . under this identification , we have @xmath223 = s_{\lambda(i)}\ ] ] where @xmath224 , and @xmath24 denotes a schur function . thus @xmath225 = s_{(0)}$ ] and @xmath226 = s_{(n - k)^k}$ ] . let @xmath227 denote the 180 degree rotation of the complement of @xmath228 inside the @xmath229 rectangle . then @xmath230 where @xmath231 . inside @xmath36 , we have the equality @xmath232 for @xmath233 . to summarize , a class @xmath234 is determined by calculating @xmath235 for all @xmath30 satisfying @xmath236 . let @xmath128 be an irreducible subvariety . recall that in section [ sec : intro ] , we defined the truncation @xmath237 ) \in h^*(\gr(k , k+m))$ ] . [ prop : main ] let @xmath238 denote the zariski - open subset of lemma [ lem : main ] for @xmath163 , and let @xmath239 where the intersection is over all @xmath240 such that @xmath130 . 1 . if @xmath237 ) = 0 $ ] then @xmath133 for all @xmath132 . if @xmath237 ) \neq 0 $ ] then for all @xmath132 , we have @xmath241 and @xmath242 = \frac{1}{d } \tau_{k+m}([w])\ ] ] where @xmath20 is the degree of @xmath140 . suppose @xmath132 . if @xmath237 ) \neq 0 $ ] then by condition ( 3 ) of lemma [ lem : main ] , we can find @xmath243}{k}$ ] satisfying @xmath130 so that @xmath244 intersects @xmath245 in a non - zero number of points . the image of these points under @xmath14 lie in @xmath136 , and since this intersection is transverse , we must have @xmath241 . for each @xmath240 , we have that @xmath124 intersects @xmath163 transversally in a finite number of points @xmath246 . also @xmath123 intersects @xmath244 transversally in a finite number of points @xmath247 , and from the conditions of lemma [ lem : main ] , we deduce that @xmath248 from lemma [ lem : inverse ] . let @xmath249 be the complement of @xmath250 in the @xmath251 rectangle . it follows from theorem [ thm : ful ] that the coefficient of @xmath252 in @xmath253 $ ] is equal to @xmath246 which is equal to @xmath254 times the coefficient of @xmath255 in @xmath212 $ ] . claim ( 2 ) follows . now suppose @xmath237 ) = 0 $ ] . then by a similar argument , we deduce that @xmath124 does not intersect any @xmath163 . this is impossible if @xmath241 since @xmath253 $ ] has a non - zero cohomology class and the intersections @xmath136 are transversal . it follows that @xmath237 ) = 0 $ ] implies that @xmath133 . claim ( 1 ) follows . let @xmath256 denote the affine coxeter group of type @xmath257 , with generators @xmath258 , and relations @xmath259 where all indices are taken modulo @xmath8 . the _ length _ @xmath260 of @xmath261 is the length of the shortest expression of @xmath262 as a product of the @xmath263 . an element @xmath264 is called _ cyclically decreasing _ if it has a reduced word @xmath265 such that @xmath266 are distinct , and if both @xmath267 and @xmath268 occur then @xmath268 occurs before @xmath267 . for example , @xmath269 is cyclically decreasing if @xmath270 . a cyclically decreasing factorization of @xmath271 is a factorization @xmath272 where @xmath273 and each @xmath274 is cyclically decreasing . for @xmath264 , we define the _ affine stanley symmetric function _ @xmath275 in @xcite it is shown that @xmath276 is a symmetric function . an affine permutation is a bijection @xmath277 satisfying 1 . @xmath278 2 . @xmath279 a _ @xmath7-bounded affine permutation _ is an affine permutation satisfying @xmath280 we denote the ( finite ) set of @xmath7-bounded affine permutations by @xmath6 . the group @xmath256 acts on the set of affine permutations on the right , with @xmath263 acting by swapping @xmath281 and @xmath282 for all @xmath283 . let @xmath284 denote the bounded affine permutation given by @xmath285 . each bounded affine permutation @xmath5 has an expression as @xmath286 for @xmath264 . the length @xmath287 of @xmath5 is declared to be equal to the length of @xmath271 . we define @xmath288 . let @xmath289 . pick a @xmath290 matrix representing @xmath103 , with columns @xmath291 , and using @xmath292 we define @xmath274 for all @xmath293 . define a function @xmath294 by @xmath295 note that if @xmath296 then @xmath297 . it is not too hard to show @xcite that @xmath298 . let @xmath299 . we define the open positroid variety @xmath300 and the positroid variety @xmath301 . we have a decomposition @xmath302 . let @xmath303 be the image of the affine stanley symmetric function in the quotient @xmath304 . [ thm : kls ] we have @xmath305 = \tf_f \in h^*(\gr(k , n))$ ] . we define the _ truncated affine stanley symmetric function _ to be @xmath306 , where @xmath9 is thought of as an element of @xmath36 . let @xmath307 be a @xmath7-bounded affine permutation , and @xmath4 be the positroid variety labeled by @xmath5 @xcite . for a general @xmath15 , we define @xmath308 to be the closure of the image of @xmath309 under @xmath14 . obviously @xmath22 depends on @xmath15 , but we will suppress this from the notation . define @xmath310 to be @xmath311 . our main result follows from theorem [ thm : kls ] and proposition [ prop : main ] applied to @xmath312 . [ thm : main ] there exists a zariski - open set @xmath313 such that 1 . if @xmath314 ) = 0 $ ] then @xmath315 for all @xmath132 , and 2 . if @xmath314 ) \neq 0 $ ] then for all @xmath132 , we have @xmath316 and @xmath317 = \frac{1}{d } \tau_{k+m}([\pi_f ] ) \in h^*(\gr(k , k+m))\ ] ] where @xmath318 is the degree of @xmath310 , which is constant for all @xmath132 . if @xmath319 for a general @xmath15 , then we declare @xmath22 to be an _ amplituhedron variety _ , and say that the affine permutation @xmath5 has _ kinematical support_. from now on , @xmath22 , @xmath310 and @xmath320 will always refer to an amplituhedron variety , the corresponding map and its degree , for @xmath132 . let @xmath5 be a @xmath7-bounded affine permutation . then @xmath5 has kinematical support if and only if for some partition @xmath228 satisfying @xmath321 , the coefficient of @xmath24 in the affine stanley symmetric function @xmath9 is non - zero . when @xmath322 , this says that @xmath5 has kinematical support if and only if @xmath323 appears in @xmath9 with non - zero coefficient . in this case , the coefficient of @xmath323 in @xmath9 can be computed using the affine pieri rule for the flag variety ( see remark [ rem : pieri ] ) . let @xmath324 , @xmath325 , and @xmath326 . bound(2,8)$ ] , which can be written as @xmath328 . then from the definitions we have @xmath329 . the coefficient of @xmath330 in @xmath9 is equal to @xmath331 . so @xmath5 has kinematical support and the map @xmath332 has degree @xmath331 . in a similar manner we can easily produce maps @xmath310 of arbitrarily high finite degree . each positroid variety @xmath4 comes from a canonical meromorphic top - form @xmath333 . if @xmath5 has kinematical support then @xmath310 is generically finite and we can define a canonical form @xmath334 by pushing forward the canonical form @xmath333 of the positroid variety . see @xcite . if @xmath325 and @xmath335 , then our notion of kinematical support essentially agrees with the notion from the theory of scattering amplitudes @xcite , though we caution that the work @xcite is mostly set in momentum space " , while the present work is set in momentum - twistor space " . physically , it is clear that when considering amplituhedron cells of dimension @xmath336 , one should restrict to those cells with kinematical support . for the cells @xmath22 of dimension @xmath337 that are used to triangulate the amplituhedron , we have an easy criterion for the degree . for @xmath338 , this is presumably the same combinatorial criterion discussed in @xcite , after translating from momentum space " to momentum - twistor space " . suppose @xmath322 . then the degree of @xmath310 is the coefficient of @xmath323 in @xmath9 , if this coefficient is positive . if this coefficient is 0 , then @xmath5 does not have kinematical support . let @xmath2 and @xmath339 be fixed , and allow @xmath8 to vary . is there a uniform bound on @xmath39 for all @xmath5 with kinematical support ? [ rem : pieri ] suppose @xmath5 has kinematical support and @xmath322 . the pieri rule for the affine flag manifold conjectured in @xcite , and proved in @xcite can be used to give a manifestly positive formula for the degree @xmath340 . specifically , the ( dual ) affine pieri rule gives an identity @xmath341 where the nonnegative numbers @xmath342 count objects called _ strong strips _ now , we have @xmath343\tf_f = [ s_{(n - k)^k } ] ( e_k)^m \tf_f\ ] ] where @xmath344\tf_f$ ] denotes the coefficient of @xmath323 when @xmath9 is expanded in terms of schur functions . by ( * ? ? ? * theorem 7.8 ) , @xmath345 \tf_g = \begin{cases } 1 & \mbox{$g \in \bound(k , n)$ and $ \ell(g ) = k(n - k)$ } \\ 0 & \mbox{otherwise . } \end{cases}\ ] ] thus @xmath346\tf_f$ ] can be obtained by counting iterated strong strips . the following result also gives a criterion for the degree of @xmath310 to be 1 . suppose @xmath5 has kinematical support and @xmath347 . let @xmath348 . then the degree of @xmath310 divides @xmath349 . in particular , if @xmath350 , then @xmath351 . for the next result , we will use the following version of zariski s main theorem . [ thm : zariski ] if @xmath64 is a quasi - compact separated scheme and @xmath352 is a separated , quasi - finite , finitely presented morphism then there is a factorization into @xmath353 , where the first map is an open immersion and the second one is finite . the following result roughly says that taking boundaries reduces the degree . so the intuition is that lower - dimensional cells tend to have smaller degree . denote by @xmath354 the _ boundary _ of a positroid variety @xmath4 . this is the union of all positroid varieties @xmath355 of strictly lower dimension . for more details on the closure partial order of positroid varieties see @xcite . suppose @xmath356 and both @xmath357 and @xmath5 have kinematical support . then @xmath358 . by applying lemma [ lem : main ] to both @xmath312 and @xmath359 , we see that we may assume that we are considering @xmath360 such that @xmath361 is dense in @xmath4 and @xmath362 is dense in @xmath363 . we may suppose that @xmath310 has degree @xmath39 and @xmath364 has degree @xmath365 where both maps are dimension - preserving . let @xmath366 consist of points @xmath367 where @xmath368 is finite . since fiber dimension is upper semicontinuous on the source , the set @xmath148 is open in @xmath369 . but @xmath370 so @xmath371 is open in @xmath362 as well . by assumption @xmath372 is quasi - finite , so by theorem [ thm : zariski ] , we have a factorization of @xmath372 as @xmath373 , where @xmath374 is an open immersion and @xmath375 is finite . clearly , @xmath375 has degree @xmath39 . it follows that the typical fiber of @xmath372 has exactly @xmath39 points , and every fiber of @xmath372 has @xmath376 points . in particular , the typical fiber of @xmath364 has @xmath376 points . thus @xmath377 . a similar argument gives suppose @xmath356 and @xmath5 does not have kinematical support . then @xmath357 does not have kinematical support . the totally nonnegative part @xmath10 @xcite of the real grassmannian is the locus of points @xmath378 representable with nonnegative ( real ) plcker coordinates @xmath379 . the totally nonnegative part of @xmath4 is defined to be @xmath380 . we say that @xmath15 is _ positive _ if all @xmath381 minors are positive , and all entries are real . if @xmath15 is positive and general ( that is , @xmath15 belongs to the zariski - dense set @xmath110 of theorem [ thm : main ] ) , we define the tnn part of @xmath22 to be @xmath382 as shown in @xcite , it is not difficult to see that in this case @xmath44 does not intersect @xmath54 . suppose @xmath5 has kinematical support . then @xmath383 . it is known @xcite that @xmath44 has real dimension equal to the complex dimension of @xmath4 . since @xmath4 is irreducible ( see @xcite ) , it follows that @xmath44 is zariski - dense in @xmath4 . it follows that @xmath384 is zariski - dense in @xmath22 and thus @xmath383 . suppose @xmath5 has kinematical support and @xmath310 has degree @xmath39 , and assume that @xmath15 is positive . while the map @xmath310 has degree @xmath39 , it is not the case that the map @xmath310 restricted to @xmath44 is generically @xmath39 to @xmath385 . it is an interesting problem to understand the geometry of the map @xmath310 when restricted to the real points @xmath386 or totally nonnegative points @xmath44 . since the truncated affine stanley symmetric function @xmath306 is schur - positive , it is also monomial - positive . however , it is not clear which monomials in the definition of @xmath9 actually contribute to the truncation . our results only apply to generic @xmath50 . however , from the point of view of convex geometry , it is interesting to consider non - generic maps . specifically , when @xmath387 , the totally nonnegative grassmannian @xmath388 is a simplex embedded in projective space , and thus _ any _ polytope @xmath389 can be expressed as the image @xmath390 for some choice of @xmath15 . when @xmath15 is positive , @xmath389 will be a cyclic polytope ( see @xcite for a related result ) .

the amplituhedron was recently introduced in the study of scattering amplitudes in @xmath0 super yang - mills . we compute the cohomology class of a tree amplituhedron subvariety of the grassmannian to be the truncation of an affine stanley symmetric function .

the cross - sections for polarised deep inelastic scattering @xcite of leptons on spin-1/2 hadrons can be expressed at leading twist as a function of three independent quark distribution functions : the unpolarised function @xmath2 , the longitudinally polarised `` helicity '' function @xmath3 and the transversely polarised or `` transversity '' function @xmath4 . this latter function is chiral - odd and therefore decouples from inclusive dis . it can however be measured in semi - inclusive dis where it appears in combination with a chiral - odd fragmentation function , the collins function @xmath5 , in an azimuthal single - spin asymmetry ( ssa ) in the hadronic end - product @xcite . a similar asymmetry could however also arise from a modulation of the transverse momentum @xmath6 of unpolarised quarks in a transversely polarised nucleon represented by the sivers function , @xmath7 @xcite . the collins and sivers effects can be disentangled in leptoproduction on a transversely polarised nucleon , since they exhibit a dependence on linearly independent kinematic variables . the collins hypothesis holds that the fragmentation function of a quark of flavour @xmath8 in a hadron @xmath9 can be written as @xcite : @xmath10 where @xmath11 is the final hadron transverse momentum with respect to the quark direction i.e. the virtual photon direction and @xmath12 is the fraction of available energy carried by the hadron ( @xmath13 is the hadron energy , and @xmath14 and @xmath15 are the incoming and scattered lepton energies respectively ) . the angle @xmath16 appearing in the fragmentation function is known as `` collins angle '' and is conveniently defined in the system where the z - axis is the virtual photon direction and the x - z plane is the muon scattering plane as @xmath17 , where @xmath18 is the hadron azimuthal angle , and @xmath19 is the azimuthal angle of the transverse spin of the struck quark ( nucleon ) . since @xmath20 , where @xmath21 is the azimuthal angle of the transverse spin of the initial quark ( nucleon ) , the relation @xmath22 is also valid . the fragmentation function @xmath23 couples to transverse spin distribution function @xmath24 and gives rise to an ssa ( denoted as @xmath25 ) dependent on the kinematic variables @xmath26 , @xmath27 and @xmath28 . following the sivers hypothesis , the difference in the probability of finding an unpolarised quark of transverse momentum @xmath29 and @xmath30 inside a polarised nucleon can be written as @xcite : @xmath31 where @xmath32 is the azimuthal angle of the quark with respect to the nucleon transverse spin orientation . it has been recently demonstrated by theoretical arguments @xcite , that the ssa ( denoted by @xmath33 ) coming from the coupling of the sivers function with the unpolarised fragmentation function @xmath34 can be observed at the leading twist from polarised semi - inclusive dis . the compass @xcite experiment on the m2 beamline of the sps accelerator at cern makes use of a high energy , intense @xmath1 beam naturally polarised by the pion - decay mechanism . the helicity of the beam is negative , and the polarisation lies at around 76% at a momentum of 160 gev / c . the spill structure of the sps in 2002 consisted of a 4.8 s burst containing approximately @xmath35 muons followed by 11.4 s out - time . the luminosity of the beam was around @xmath36 @xmath37s@xmath38 . the spectrometer itself comprises two stages each equipped with a spectrometer magnet and detectors for tracking , energy determination and particle identification @xcite . compass uses the polarised target system of the smc experiment which consists of two @xmath39lid cells , each 60 cm long and with a radius of 1.5 cm located one after the other along the beam axis . in longitudinal running a 2.5 t field is maintained by a solenoid magnet , supplemented in transverse running by an orthogonal 0.5 t dipole field . the relaxation time of around 2000 hours of the dipole field is sufficient to allow the transverse measurement . the two target cells are always maintained with opposite polarisation . transverse polarisations @xmath40 of + /- 50 % are routinely reached ; the dilution factor @xmath41 of the target material is calculated to be around 0.38 . two periods of data with transverse target polarisation were taken in 2002 . in order to compensate the different acceptances of the target cells , their polarity was flipped by microwave reversal halfway through each period . the counting rate asymmetry between these two sub - periods was then measured for each target - cell separately . the number of events in each of the two polarisation states ( @xmath42/@xmath43 ) may be written as a function of the collins ( sivers ) angle as @xmath44 where @xmath45 is the amplitude of the experimental asymmetry and @xmath46 is a function containing the acceptance of the apparatus . the angles are always calculated assuming the target polarisation point up in the lab system . the `` raw '' amplitude @xmath45 is connected to the collins ( sivers ) asymmetry through the expression @xmath47 with @xmath48 is the spin transfer coefficient or `` depolarisation factor '' . indicating with @xmath49 the fractional energy transfer from muon to virtual photon in the initial scattering process , @xmath48 may be calculated from the kinematics of each individual event through @xmath50 in the case of the collins effect . for the sivers effect @xmath51 , since this effect deals with unpolarised quarks . in each period @xmath52 ( @xmath53 ) is fitted separately for the two target cells from the event flux with the two target orientations using the expression @xmath54 where @xmath55 is the ratio of the total number of events in the two target polarisation orientations . the total sample collected in the 2002 transverse data - taking periods amounted to about 200 pb@xmath38 in terms of integrated luminosity . events were selected in which a primary vertex with identified beam and scattered muon and at least one outgoing hadron was found in one of the two target cells . a clean separation of muon and hadron samples was achieved by cuts on the amount of material traversed in the spectrometer . in addition , the kinematic cuts @xmath56(g@xmath57v@xmath58 , @xmath59 g@xmath57v/@xmath60 and @xmath61 were applied to the data to ensure a deep - inelastic sample above the region of the nuclear resonances and within the compass trigger acceptance . the upper bound on @xmath49 also served to keep radiative corrections small . the asymmetries were calculated for two different samples : both for all hadrons emerging from the primary vertex , and , a sub - sample of this , only for the most energetic or `` leading '' hadron in each event . the leading hadron was determined as the most energetic non - muonic particle of the primary vertex with an energy component @xmath62 and a transverse momentum @xmath63 g@xmath57v/@xmath64 . when all the hadrons coming from the primary vertex were considered , the @xmath27 cut was lowered to @xmath65 . the final data sample had average values @xmath66 , @xmath67 and @xmath68 ( g@xmath57v@xmath58 . the average value for @xmath27 and @xmath69 are @xmath70 and @xmath71 g@xmath57v/@xmath64 respectively for the leading hadron analysis , and @xmath72 and @xmath73 g@xmath57v@xmath74 in the all hadrons case . , @xmath27 and @xmath28 . leading hadrons analysis on top canvas , all hadrons on bottom canvas.,title="fig : " ] , @xmath27 and @xmath28 . leading hadrons analysis on top canvas , all hadrons on bottom canvas.,title="fig : " ] the results of the asymmetries plotted against the kinematic variables @xmath26 , @xmath27 and @xmath28 are shown in fig . [ results ] for positive ( full points ) and negative ( open points ) hadrons , and for the analysis of leading ( top plot ) hadron and all hadrons ( bottom plot ) . stability checks showed that the ratio of acceptances and efficiencies as a function of the collins ( sivers ) angle does not change between two spin orientations . furthermore the results were stable when each target cell was split into two parts and the asymmetry calculated separately , and when the data was split into high and low hadron momenta . this leads to the conclusion that systematic effects are smaller than the statistical errors . in these first measurements of transverse spin effects on a deuteron target , the collins and sivers asymmetries were found to be compatible with zero within the statistical accuracy of the data . a marginal indication of a collins effect at large @xmath27 for both positive and negative charges is seen . coupled with hermes results on a proton target also reported at this conference , this may imply a negative neutron contribution to the asymmetries . a sensitivity improvement of factor two is expected once the data from the 2003 and 2004 compass runs have been analysed . from 2006 onwards , a target magnet with a significantly larger acceptance should increase the statistics at high @xmath26 . complementary measurements with a proton target are also planned . 9 v. barone , a. drago and p. g. ratcliffe , phys . * 359 * ( 2002 ) j. c. collins , nucl . b * 396 * ( 1993 ) 161 . d. w. sivers , phys . d * 41 * ( 1990 ) 83 . [ compass collaboration ] , cern - spslc-96 - 14 . m. anselmino , v. barone , a. drago , f. murgia , [ arxiv : hep - ph/0209073 ] . s. j. brodsky , d.s . hwang , i. schmidt , phys . b * 530 * ( 2002 ) 99 . j. c. collins , phys . b * 536 * ( 2002 ) 43 . et al . _ , `` recent compass spin physics results '' , these proceedings . mallot , nim * 518*(2004 ) 121 .

compass is a fixed - target experiment on the sps m2 beamline at cern . its @xmath0lid target can be polarised both longitudinally and transversally with respect to the longitudinally polarised 160gev / c @xmath1 beam . approximately 20% of the beam - time in 2002 , 2003 and 2004 was spent in the transverse configuration , allowing the first measurement of both the collins and sivers asymmetries on a deuterium target . first results from the the transverse data of the compass run in 2002 are reported here .

effect of chemical reactions on turbulent transport is of great importance in many applications ranging from atmospheric turbulence and transport of pollutants to combustion processes ( see , e.g. , @xcite ) . even without turbulence , the combustion is an intrinsically complex process with chemical kinetic mechanism which may involve hundreds species and up to thousands chemical reactions . in most technological applications ( e.g. spark - ignition engines , industrial burners , furnaces etc . ) combustion nearly always occurs within a turbulent rather than a laminar flow . when a flame propagates in a turbulent flow , turbulence enhances mixing of fuel and oxidizer and therefore enhances combustion . the combustion heat in turn generates flow instabilities and affects the turbulence and turbulent transport . it has been noticed @xcite that the latter can manifest themselves in a form of the so - called counter - gradient scalar transport , i.e. a flux of products from unburnt to burnt regions of a premixed flame . in its turn , the counter - gradient transport can substantially reduce the flame speed @xcite . thus , understanding the turbulent transport in chemically reacting flows is of great importance for calculations of the turbulent flame speed and plays a key role in the premixed turbulent combustion . however , besides a few studies @xcite , the vast majority of available models of the influence of combustion on turbulent transport @xcite are focused solely on effects that stem from density variations , i.e. , neglecting straightforward influence of chemical reactions on turbulent fluxes @xcite . the effect of chemical reactions on turbulent diffusion of chemically reacting gaseous admixtures in a developed turbulence has been studied analytically using a path - integral approach for a delta - correlated in time random velocity field @xcite . this phenomenon also has been recently investigated applying the spectral - tau approach that is valid for large reynolds and peclet numbers @xcite . these studies have demonstrated that turbulent diffusion of the reacting species can be strongly suppressed by a large factor ( of the order of the turbulent damkhler number , @xmath3 , that is a ratio of turbulent , @xmath4 , and chemical , @xmath5 , time scales ) . the dependence of turbulent diffusion coefficient , @xmath6 , versus the turbulent damkhler number obtained theoretically in @xcite , was validated using results of mean - field simulations of a reactive front propagating in a turbulent flow @xcite . in these simulations , the mean speed , @xmath7 , of the planar one - dimensional reactive front was determined using numerical solution of the kolmogorov - petrovskii - piskunov ( kpp ) equation @xcite or the fisher equation @xcite . this equation was extended in @xcite to take into account memory effects of turbulent diffusion when the turbulent time was much larger than the characteristic chemical time . turbulent diffusion coefficients as a function of @xmath1 were determined numerically in @xcite using the obtained function @xmath8 and invoking the well - known kpp expression , @xmath9 . the theoretical dependence @xmath10 derived in @xcite was in a good agreement with the numerical results of @xcite . the goal of the present study is to further validate the theory of the turbulent diffusion of chemically reacting flows developed in @xcite , by comparing its predictions with the three - dimensional direct numerical simulations ( dns ) of the reaction wave propagating in a homogeneous isotropic and incompressible turbulence for a wide range of ratios of the wave speed to the r.m.s . turbulent velocity and different reynolds numbers . this paper is organised as follows . the theoretical results @xcite to be tested are summarized in section ii . dns performed to validate the theory are described in section iii . subsequently , validation results are discussed in section iv , and concluding remarks are given in section v. equation for the scalar field in the incompressible chemically reacting turbulent flow reads : @xmath11 where @xmath12 is a scalar field , @xmath13 is the instantaneous fluid velocity field , @xmath14 is the diffusion coefficient based on molecular fick s law , @xmath15 is the source ( or sink ) term , @xmath16 is the stoichiometric coefficient that is the order of the reaction with respect to species @xmath17 , @xmath18 is the overall order of the reaction , and @xmath19 is total number of species . the function @xmath20 is usually chosen according to the arrhenius law ( to be given in the next section ) . the velocity @xmath21 of the fluid is determined by the navier - stokes equation : @xmath22 where @xmath23 is the external force to support turbulence , @xmath24 is the kinematic viscosity , @xmath25 and @xmath26 are the fluid pressure and density , respectively . for an incompressible flow the fluid density is constant . to determine turbulent transport coefficients , eq . ( [ a1 ] ) is averaged over an ensemble of turbulent velocity fields . in the framework of a mean - field approach , the scalar field @xmath27 is decomposed into the mean field , @xmath28 , and fluctuations , @xmath29 , where @xmath30 , and angular brackets imply the averaging over the statistics of turbulent velocity field . the velocity field is decomposed in a similar fashion , @xmath31 , assuming for simplicity vanishing mean fluid velocity , @xmath32 , where @xmath33 are the fluid velocity fluctuations . the equation for the turbulent flux , @xmath34 , derived in @xcite , is given by : @xmath35 where @xmath36 is the coefficient of turbulent diffusion of the scalar field , @xmath37 @xmath38 is the non - dimensional function , @xmath39 and @xmath40 , @xmath41 is the characteristic turbulent velocity in the integral scale , @xmath42 , of turbulence , @xmath43 is the characteristic turbulent time . we define the turbulent damkhler number , @xmath3 , as the ratio of turbulent time in the integral scale and the chemical time , @xmath44 , where @xmath45 is the mean source function , and @xmath46 the term @xmath47 in the expression ( [ a11 ] ) for the turbulent flux of reacting admixtures , @xmath48 , describes the mutual turbulent diffusion of admixtures , where @xmath49 is the coefficient of the mutual turbulent diffusion of admixtures , @xmath50 it is known in the irreversible thermodynamics that the mutual molecular diffusion of admixtures is caused by interaction between gaseous admixtures due to collisions of molecules of the admixtures . in turbulent flow with chemical reactions inhomogeneities of the number density of one of the reagents causes fast change ( during the chemical reaction time scale @xmath51 ) of the number density of other components due to the shift from the chemical equilibrium . the asymptotic behavior of the function @xmath38 is as follows . when @xmath52 @xmath53 while for @xmath54 , @xmath55 the stoichiometric coefficient @xmath16 in eq . ( [ a1 ] ) is known as the order of the reaction with respect to species @xmath17 . in practice the overall order of the reaction is defined as the sum of the exponents of the concentrations in the reaction rate , @xmath56 . for a simplified model of a single - step reaction typically used in numerical simulations of turbulent combustion , the overall order of the reaction is the molecularity of the reaction , indicating the number of particles entering the reaction . for example , for the first - order chemical reaction @xmath57 , and the coefficient of turbulent diffusion , @xmath0 , of the scalar field is : @xmath58 so that for @xmath52 , @xmath59 while for @xmath54 , @xmath60 for the first - order reaction , @xmath61 , it follows from eq . ( [ aaa17 ] ) that turbulent diffusion of the number density of admixtures for a large turbulent damkhler number , @xmath62 is strongly reduced , i.e. , @xmath63 . this implies that the turbulent diffusion of admixture for a large turbulent damkhler number is determined by the chemical time . the underlying physics of the strong reduction of turbulent diffusion is quite transparent . for a simple first - order chemical reaction @xmath61 the species @xmath64 of the reactive admixture are consumed and their concentration decreases much faster during the chemical reaction , so that the usual turbulent diffusion based on the turbulent time @xmath65 , does not contribute to the mass flux of a reagent @xmath64 . the turbulent diffusion during the turnover time of the turbulent eddies is effective only for the product of reaction , @xmath66 . dns of a finite thickness reaction wave propagation in forced , homogeneous , isotropic , and incompressible turbulence for the first - order chemical reactions , were performed in a fully periodic rectangular box of size of @xmath67 using a uniform rectangular mesh of @xmath68 points and a simplified in - house solver @xcite developed for low - mach - number reacting flows . contrary to recent dns studies by two of the present authors @xcite that addressed self - propagation of an infinitely thin interface by solving a level set equation , the present simulations deals with a wave of a finite thickness , modelled with eq . ( [ a1 ] ) for a single combustion progress variable @xmath69 ( @xmath70 and 1 in unburnt and burnt mixtures , respectively ) , while the navier - stokes equation ( [ a2 ] ) was numerically integrated in both cases . to mimic a highly non - linear dependence of the reaction rate @xmath71 on @xmath69 in a typical premixed flame characterized by significant variations in the density @xmath26 and temperature @xmath72 , the following expression @xmath73 \label{a18}\end{aligned}\ ] ] was invoked in the present constant - density simulations . here , @xmath74 is a reaction time scale , while parameters @xmath75 and @xmath76 are counterparts of the zeldovich number @xmath77 and heat - release factor @xmath78 , respectively , which are widely used in the combustion theory @xcite , with subscripts @xmath79 and @xmath80 designating unburned and burned mixtures , respectively . indeed , substitution of @xmath81 into exponent in eq . ( [ a18 ] ) results in the classical arrhenius law @xmath82 i.e. eq . ( [ a18 ] ) does allow us to mimic behavior of reaction rate in a flame by considering constant - density reacting flows . the speed @xmath83 of the propagation of the reaction wave in the laminar flow , the wave thickness @xmath84 , the wave time scale @xmath85 , and the damkhler number used in dns , @xmath86 , were varied by changing the diffusion coefficient @xmath87 and the reaction time scale @xmath74 , which were constant input parameter for each dns run . the present dns are similar to dns discussed in detail in @xcite , except for substitution of a level set equations used in @xcite by eqs . ( [ a1 ] ) and ( [ a18 ] ) . therefore , we will restrict ourselves to a very brief summary of the simulations . the boundary conditions were periodic not only in transverse directions @xmath88 and @xmath89 , but also in direction @xmath90 normal to the mean wave surface . in other words , when the reaction wave reached the left boundary ( @xmath91 ) of the computational domain , the identical reaction wave entered the domain through its right boundary ( @xmath92 ) . the initial turbulence field was generated by synthesizing prescribed fourier waves @xcite with an initial rms velocity @xmath93 and the forcing scale @xmath94 . subsequently , a forcing function @xmath23 , see eq . ( [ a2 ] ) , was invoked in order to maintain statically stationary turbulence following the method described in ref . @xcite . as shown earlier @xcite , ( i ) the rms velocity @xmath93 was maintained as the initial value , ( ii ) the normalized dissipation rate @xmath95 averaged over the computational domain fluctuated slightly above 3/2 after a short period ( @xmath96 ) of rapid transition from the initial artificially synthesized flow to the fully developed turbulence , ( iii ) the forced turbulence achieved good statistical homogeneity and isotropy over the entire domain , and ( iv ) the energy spectrum showed a sufficiently wide range of the kolmogorov scaling ( @xmath97 ) at the reynolds number , @xmath98 , based on the scale @xmath99 . three turbulent fields were generated by specifying three different initial turbulent reynolds numbers @xmath100 , 100 , and 200 , which were increased by increasing the domain size @xmath101 . the increase in the width @xmath101 of the computational domain resulted in increasing the integral scale @xmath42 , the turbulent time scale @xmath4 , and , hence , the damkhler number @xmath102 . the simulation parameters are shown in table i. because a reaction wave does not affect turbulence in the case of constant density @xmath26 and viscosity @xmath24 , the flow statistics were the same in all cases that had different @xmath83 , but the same @xmath103 . it is worth noting that the length scale @xmath42 reported in table i and used to evaluate @xmath102 was averaged over the computational domain and time at @xmath104 and was lower than its initial value @xmath94 by 10 or 20 % at @xmath105 or 200 , respectively @xcite . .dns cases . [ cols="^,^,^,^,^",options="header " , ] dependencies of the normalized turbulent scalar flux @xmath106 on the mean combustion progress variable @xmath107 , computed for ( a ) @xmath100 , ( b ) @xmath105 , and ( c ) @xmath108.,title="fig:",width=294 ] dependencies of the normalized turbulent scalar flux @xmath106 on the mean combustion progress variable @xmath107 , computed for ( a ) @xmath100 , ( b ) @xmath105 , and ( c ) @xmath108.,title="fig:",width=294 ] dependencies of the normalized turbulent scalar flux @xmath106 on the mean combustion progress variable @xmath107 , computed for ( a ) @xmath100 , ( b ) @xmath105 , and ( c ) @xmath108.,title="fig:",width=294 ] when the width @xmath101 was increased by a factor of two , the numbers @xmath109 , @xmath110 , and @xmath111 were also increased by a factor of two , i.e. @xmath112 , 512 , or 1024 at @xmath100 , 100 , or 200 , respectively . accordingly , in all cases , the kolmogorov length scale @xmath113 evaluated using the dissipation rate @xmath114 averaged over volume ( angle brackets ) and time at @xmath115 ( overbars ) was of the order of the grid cell size @xmath116 , thus , indicating sufficient grid resolution . capability of the used grids for well resolving not only the kolmogorov eddies , but also the reaction wave was confirmed in separate ( i ) 1d simulations of planar laminar flames and ( ii ) 2d simulations @xcite of laminar flames subject to hydrodynamic instability @xcite . moreover , the resolution of the present dns was validated by running simulations with the grid cell size @xmath116 decreased by a factor of four at @xmath100 , i.e. by setting @xmath109 equal to 1028 . in the next section , we will report the mean quantities @xmath117 averaged over a transverse plane of @xmath118const and time at @xmath119 , with @xmath120 being equal to @xmath121 or even longer . moreover , we will present correlations between fluctuating quantities @xmath122 . furthermore , using the computed dependencies of @xmath123 , the dependencies of other mean quantities and correlations on distance @xmath90 will be transformed to dependencies of these variables and correlations , respectively , on the mean combustion progress variable @xmath107 . figure [ fig1 ] shows dependencies of the normalized turbulent scalar flux @xmath106 versus the mean combustion progress variable @xmath107 , computed for ( a ) @xmath100 , ( b ) @xmath105 , and ( c ) @xmath108 . while , the flux magnitude depends on @xmath124 and , hence , on @xmath125 , see table i , such variations in the flux magnitude are sufficiently weak and non - monotonic , with the peak magnitude being obtained at a medium @xmath126 . dependencies of the normalized mean turbulent wave thickness @xmath127 on the normalized wave speed @xmath124 computed at three different turbulent reynolds numbers , @xmath103 , specified in legends . , width=294 ] on the contrary , the mean turbulent wave thickness @xmath128 defined using the maximum gradient method , i.e. , @xmath129 decreases rapidly with the increase of the normalized wave speed @xmath124 and , hence , @xmath125 , see fig . [ fig2 ] . accordingly , the gradient of the mean combustion progress variable is increased by @xmath124 ( or @xmath125 ) , whereas turbulent diffusivity evaluated as follows , @xmath130 is decreased with increasing @xmath124 and @xmath125 , see fig . [ fig3 ] . the decrease of the turbulent diffusion coefficient , @xmath0 , with the increase of the damkhler number observed in dns , agrees qualitatively with the theory . dependencies of the normalized mean turbulent scalar diffusivity @xmath131 on the mean combustion progress variable @xmath107 , computed for ( a ) @xmath100 , ( b ) @xmath105 , and ( c ) @xmath108 . values of @xmath124 are specified near curves . , title="fig:",width=283 ] dependencies of the normalized mean turbulent scalar diffusivity @xmath131 on the mean combustion progress variable @xmath107 , computed for ( a ) @xmath100 , ( b ) @xmath105 , and ( c ) @xmath108 . values of @xmath124 are specified near curves . , title="fig:",width=283 ] dependencies of the normalized mean turbulent scalar diffusivity @xmath131 on the mean combustion progress variable @xmath107 , computed for ( a ) @xmath100 , ( b ) @xmath105 , and ( c ) @xmath108 . values of @xmath124 are specified near curves . , title="fig:",width=283 ] moreover , fig . [ fig3 ] indicates that @xmath0 evaluated using eq . ( [ edt ] ) depends weakly on @xmath107 , thus , implying that the influence of the reaction on the turbulent diffusion coefficient may be characterized with a single mean turbulent diffusivity defined as follows @xmath132 values of the mean turbulent diffusion coefficient , @xmath133 , obtained in the simulations ( symbols ) are compared with the theoretical predictions for @xmath0 in fig . [ fig4]a , with the curve being calculated substituting @xmath1 in eq . ( [ a17 ] ) with @xmath125 . the functional dependence of @xmath133 on @xmath1 is the same as predicted by theory . the dns data show a substantial decrease in @xmath133 with increasing @xmath125 , but the dns data are shifted to larger damkhler numbers when compared to the theoretical curve . such a quantitative difference between the theory and simulations is not surprising bearing in mind different definitions of damkhler numbers in the theory and simulations . validation of the theory . symbols show dns data , with the reynolds numbers @xmath103 being specified in legends . curves have been computed using eq . ( [ a17 ] ) . ( a ) dns data are normalized using @xmath134 . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath125 . ( b ) dns data are normalized using the maximum ( for various @xmath135 ) turbulent diffusivity @xmath133 computed for each @xmath103 independently . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath136 , with the tuned values of the re - scaling factor @xmath80 being specified in legends . ( c ) dns data are normalized using @xmath137 , with the tuned values of the re - scaling factor @xmath138 being specified in legends . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath136 , with the tuned values of the re - scaling factor @xmath80 being specified in legends . , title="fig:",width=283 ] validation of the theory . symbols show dns data , with the reynolds numbers @xmath103 being specified in legends . curves have been computed using eq . ( [ a17 ] ) . ( a ) dns data are normalized using @xmath134 . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath125 . ( b ) dns data are normalized using the maximum ( for various @xmath135 ) turbulent diffusivity @xmath133 computed for each @xmath103 independently . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath136 , with the tuned values of the re - scaling factor @xmath80 being specified in legends . ( c ) dns data are normalized using @xmath137 , with the tuned values of the re - scaling factor @xmath138 being specified in legends . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath136 , with the tuned values of the re - scaling factor @xmath80 being specified in legends . , title="fig:",width=283 ] validation of the theory . symbols show dns data , with the reynolds numbers @xmath103 being specified in legends . curves have been computed using eq . ( [ a17 ] ) . ( a ) dns data are normalized using @xmath134 . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath125 . ( b ) dns data are normalized using the maximum ( for various @xmath135 ) turbulent diffusivity @xmath133 computed for each @xmath103 independently . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath136 , with the tuned values of the re - scaling factor @xmath80 being specified in legends . ( c ) dns data are normalized using @xmath137 , with the tuned values of the re - scaling factor @xmath138 being specified in legends . in eq . ( [ a17 ] ) , @xmath1 is substituted with @xmath136 , with the tuned values of the re - scaling factor @xmath80 being specified in legends . , title="fig:",width=283 ] it is worth checking whether or not the observed quantitative difference can be reduced by re - scaling one of the two damkhler numbers . accordingly , for each @xmath103 independently , we ( i ) substituted @xmath1 with @xmath136 in eq . ( [ a17 ] ) , ( ii ) varied the re - scaling factor @xmath139 from 1.0 to 0.001 with a step of 0.001 , and ( iii ) selected the factor that yielded the least rms difference @xmath140 between eq . ( [ a17 ] ) and the dns data . note that , because ( i ) different definitions of the length scale @xmath42 were used in the the theory and simulations and ( ii ) the asymptotic diffusion coefficient @xmath141 is beyond the scope of the present study , which aims at the reduction in turbulent diffusivity due to chemical reactions , the dns data were normalized using the maximum ( for various @xmath135 ) turbulent diffusivity @xmath0 computed for each @xmath103 . figure [ fig4]b shows that such a re - scaling transformation substantially improves agreement between the theory and simulations . the fact that the tuned values of the re - scaling factor @xmath80 , reported in legends , differ significantly from unity and depend on @xmath103 is not surprising . indeed , as discussed in section ii.b , the theoretical damkhler number , @xmath1 , is defined using the mean reaction rate @xmath142 , whereas @xmath136 is inversely proportional to @xmath143 which characterizes the peak reaction rate @xmath144 in the laminar wave . in other words , @xmath145 , but the latter ratio is well - known to be much less than unity , e.g. in turbulent flames , due to ( i ) highly non - linear dependence of @xmath71 on @xmath69 and ( ii ) significant magnitude of fluctuations of @xmath146 . therefore , large values of @xmath80 reported in legends in fig . [ fig4]b are expected . an increase in the reported values of @xmath80 by @xmath103 is also expected , because an increase in @xmath147 results in decreasing probability of finding intermediate states ( @xmath148 ) of the mixture , thus , making the local record of @xmath146 more close to random telegraph signal and increasing the magnitude of fluctuations in @xmath146 . it is worth noting , however , that the use of @xmath149 for normalization of the dns data does not seem to be the best solution , because the data do not indicate saturation at a level of @xmath150 . accordingly , the following test was also performed . in addition to substitution of @xmath151 with @xmath136 in eq . ( [ a17 ] ) , as discussed before , the mean diffusivity @xmath133 was normalized using @xmath137 . by varying not only @xmath139 ( from 1.0 to 0.001 with a step of 0.001 ) , but also another re - scaling factor @xmath138 ( from 1.0 to 1.5 with a step of 0.01 ) , the optimum pair of the re - scaling factors was selected by minimizing the following rms difference @xmath152 as shown in fig . [ fig4]c , such a double re - scaling transformation allowed us to obtain excellent quantitative agreement between the theory and the present dns . due to the use of two tuning parameters @xmath80 and @xmath138 , this quantitative agreement should not be overestimated and the results of the performed tests should be considered to be qualitative validation . however , it is worth noting that there is a large number of theoretical expressions that involve two tuning parameters , but do not allow to obtain a quantitative agreement with experimental data . for instance , in the field of premixed combustion , the following theoretical expression @xmath153 is widely used to evaluate two unknown quantities , the unperturbed laminar flame speed @xmath154 and markstein number , @xmath155 , by approximating measured dependencies of perturbed laminar flame speed @xmath83 by stretch rate @xmath156 . nevertheless , as reviewed elsewhere @xcite and shown in recent papers @xcite , there are a number of experimental curves @xmath157 that are poorly fitted by this expression in spite of tuning @xmath154 and @xmath155 . validation of the recent theory @xcite of the effect of chemical reaction on turbulent diffusion using three - dimensional dns of a finite thickness reaction wave propagation in forced , homogeneous , isotropic , and incompressible turbulence for the first - order chemical reactions , has revealed a very good qualitative agreement between the theoretical predictions and the dns results . in particular , an increase in damkhler number results in strong reduction of turbulent diffusion coefficient in the theory and dns . this research was supported in part by the israel science foundation governed by the israeli academy of sciences ( grant no . 1210/15 , te , nk , ir ) , the research council of norway under the frinatek ( grant 231444 , nk , ml , ir ) , the swedish research council ( ry ) , the chalmers combustion engine research center ( cerc ) and chalmers transport and energy areas of advance ( al ) . this research was initiated during nordita program _ physics of turbulent combustion _ , stockholm ( september 2016 ) . the computations were performed on resources provided by the swedish national infrastructure for computing ( snic ) at beskow - pdc center .

three - dimensional direct numerical simulations ( dns ) of a finite thickness reaction wave propagation in forced , homogeneous , isotropic , and incompressible turbulence for the first - order chemical reactions have been performed to validate the theoretically predicted effect of chemical reactions on turbulent diffusion . we show that the dns data agree with the theory at least qualitatively . in particular , in line with the theoretical predictions , the dns data show a significant decrease in the turbulent diffusion coefficient @xmath0 with increasing damkhler number @xmath1 , which characterizes a ratio of turbulent and reaction time scales . quantitative differences are attributed to different definitions of the damkhler number and turbulence length scale , adopted in the theory and simulations . re - scaling of these two variables results in excellent quantitative agreement between predicted and computed functions @xmath2 .

non - trivial power laws are known to characterize second order phase transitions . a great success of the theory of critical phenomena has been to develop methods allowing to predict these power laws @xcite . one of the most successful approaches used in the theory of critical phenomena is the renormalization group , which consists in trying to relate physical properties of a given system at different values of the external parameters ( like the temperature or the magnetic field ) . in the last three or four decades , other non - trivial power laws @xcite have been found in all kinds of systems : transition to chaos by period doubling @xcite , geometrical problems like self avoiding walks ( which model polymers ) and random walks @xcite , sand pile models and several other self organised critical systems @xcite , coarsening @xcite , etc . in many cases , renormalization ideas could be extended to predict the exponents of these power laws . in this work , we report recent results on simple models of genealogical trees @xcite . when one looks at the distribution of repetitions in a genealogical tree ( in the framework of the simple models defined below ) , one observes non trivial power laws . the exponents of these power laws can be calculated _ exactly _ by writing a relation on the generating function of the weights of the ancestors ( a quantity proportional to the number of times they appear in a genealogical tree ) which has the form of a simple renormalization transformation . beyond the intrinsic interest of these models to describe real genealogies , they constitute simple pedagogical examples for which renormalization ideas allow the exact prediction of non trivial exponents . let us first consider a simple neutral model of a closed population with sexual reproduction . by definition of the model , the population size at generation @xmath0 in the past is @xmath6 and each individual at generation @xmath0 has two parents chosen at random among the @xmath7 individuals in the previous generation @xmath8 . here @xmath0 counts the number of past generations and so increases as one climbs up a genealogical tree . for simplicity we will consider either a population of constant size ( @xmath9 ) or a population size increasing exponentially with an average number @xmath10 of offsprings per couple , i.e. as @xmath0 counts the number of past generations ; @xmath11 is the size of the population at present , while the constant size case corresponds to @xmath12 . a related model was introduced to study the genetic similarity between individuals in a population evolving under sexual reproduction @xcite , although there the two parents were distinct . we do not exclude this case here . clearly , the number of branches of the genealogical tree of any individual increases like @xmath13 and , as soon as the number of branches exceeds @xmath6 , there should be repetitions in this tree . let us denote by @xmath14 the number of times that an individual @xmath15 living at generation @xmath0 in the past appears in the genealogical tree of individual @xmath16 . at generation @xmath17 , the only individual in the tree of @xmath16 is @xmath16 itself , therefore @xmath18 and the evolution of these repetitions satisfies the recursion @xmath19 the quantity we want to consider is the probability @xmath20 that an individual living at generation @xmath0 in the past appears @xmath21 times in the genealogical tree of individual @xmath16 ( living at generation @xmath22 ) . normalization implies @xmath23 the initial condition ( [ initial ] ) gives @xmath24 and the fact that each individual has two parents at the previous generation gives @xmath25 of @xmath21 repetitions after @xmath0 generations ( @xmath26 is not shown ) at @xmath27 and @xmath28 for a population of constant size . in figure 1a , @xmath29 and in figure 1b , @xmath30 . both figures show averages over 1000 samples.,width=340 ] of @xmath21 repetitions after @xmath0 generations ( @xmath26 is not shown ) at @xmath27 and @xmath28 for a population of constant size . in figure 1a , @xmath29 and in figure 1b , @xmath30 . both figures show averages over 1000 samples.,width=340 ] these probabilities @xmath20 can be measured by simulating small systems through a monte carlo procedure : for each individual of a population at generation @xmath0 , two parents are chosen at random among the @xmath7 individuals at generation @xmath8 . figure 1 shows the results of such simulations for two populations of constant sizes , @xmath31 for several values of @xmath0 with @xmath32 in fig . 1a and @xmath33 in fig . we see that for small @xmath0 there are very few repetitions and @xmath20 decreases very fast with @xmath21 . on the other hand , when @xmath0 increases , the shape of @xmath20 becomes independent of @xmath0 and of the population size @xmath34 , with a clear power law at small @xmath21 and a fast decay at large @xmath21 . figure 2 shows the distribution @xmath20 for several values of @xmath0 and a population which increases exponentially with time , @xmath35 . here , again , the shape becomes stationary in the interval where @xmath0 is large enough and @xmath6 is still large . this stationary shape is different from the one seen in fig . 1 . for a population size increasing by a factor @xmath36 at each generation . here @xmath37 , and averages over @xmath38 samples are performed . the generations shown are @xmath39 and @xmath40.,width=340 ] the shape of @xmath20 becomes stationary for large @xmath6 and large @xmath0 in the sense that one gets a fixed distribution by an appropriate rescaling . in fact , introducing the rescaled quantities @xmath1 and @xmath2 @xmath41 @xmath42 where @xmath1 can be considered as a continuous variable for @xmath43 , ( [ norm1],[norm2 ] ) transform into @xmath44 and we expect @xmath2 to become a fixed distribution @xmath3 . this means that if we associate to each individual @xmath15 in the tree of @xmath16 at generation @xmath0 in the past a weight defined by @xmath45 the distribution of these weights becomes stationary in the scaling limit . from ( [ evolution],[wi ] ) it is clear that these weights satisfy @xmath46 as we limit ourselves to the case of a population increasing exponentially at rate @xmath10 per generation ( so that @xmath47 ) , ( [ evolutionw ] ) reduces to @xmath48 the ratios @xmath49 can be interpreted as the probability of reaching individual @xmath15 by randomly climbing up the genealogical tree of @xmath16 . in the particular case of a population of constant size ( @xmath12 ) , the factor @xmath50 in ( [ wp ] ) is easy to understand . for a population of increasing size ( @xmath51 ) , there is a factor @xmath52 in ( [ wp ] ) instead of @xmath50 because of the factor @xmath6 in the definition ( [ wi ] ) of the weights @xmath53 . the key observation which allows one to calculate the distribution @xmath2 in the scaling limit ( large @xmath0 and large @xmath6 ) is that , for large @xmath6 and for large @xmath0 , _ the random variables @xmath54 which appear in the r.h.s . of ( [ wp ] ) become independent_. this is due to the fact that ( at least in the model we consider ) the weights @xmath55 ( of brothers and sisters ) in the r.h.s . of ( [ wp ] ) are uncorrelated . this independence , which is discussed in the appendix , will be the basis of the calculation of the fixed distribution @xmath3 in the following sections . one can consider some variants of the model defined above , for instance : * at each generation one could form fixed couples by making random pairs and assign to each individual at generation @xmath0 one of these pairs ( of parents ) chosen at random at the previous generation ( @xmath8 ) . in this case the correlations between the weights @xmath56 would again be small in the scaling limit and they can be ignored in the r.h.s . of ( [ wp ] ) . * one can also consider an imaginary situation where each individual has @xmath57 parents ( instead of @xmath58 ) . in this case , the definition of the weights ( [ wi ] ) should be replaced by @xmath59 to keep @xmath2 normalized as in ( [ norm ] ) . for a population of constant size @xmath60 , the evolution of the weights ( [ wp ] ) becomes @xmath61 as shown in the appendix , in the scaling limit , the correlations on the r.h.s . of ( [ wpp ] ) can be neglected in this case too . in the remaining of this work , we try to predict the stationary shape @xmath3 . the fact that the weights in the r.h.s . of ( [ wp ] ) are uncorrelated greatly simplifies the problem . one can then consider that @xmath62 is the sum of @xmath63 independent identically distributed random variables @xmath55 , where @xmath63 is itself random . the probability @xmath64 of @xmath63 is clearly @xmath65 which for large @xmath6 becomes ( using the fact that @xmath66 ) a poisson distribution @xmath67 therefore for large @xmath6 , the number @xmath63 of terms ( @xmath63 is the number of children of @xmath15 ) in the r.h.s . of ( [ wp ] ) is randomly distributed according to ( [ poisson ] ) and these @xmath63 terms are uncorrelated . this becomes a problem of branching processes @xcite . if one introduces the generating function @xmath68 @xmath69 \rangle \label{qdef}\ ] ] and uses ( [ wp ] ) and the fact that the weights are independent , one finds that @xmath68 satisfies @xmath70 . \label{qrecursion}\ ] ] the normalization ( [ wi ] ) of the @xmath71 implies that we have for all @xmath0 @xmath72 recursions similar to ( [ qrecursion ] ) appear in the theory of branching processes , in particular in the galton - watson process , already introduced in the 19th century to study the problem of the extinction of families @xcite . from ( [ qdef],[qrecursion ] ) , one can easily obtain recursions for the moments of the weigths @xmath53 , @xmath73 and so on . we see that for large @xmath0 , each moment of @xmath74 has a limiting value , as expected from the observation in the previous section that @xmath2 converges to a fixed distribution @xmath3 such that @xmath75 the limiting values of these moments @xmath76 etc . , can be obtained directly by expanding the solution @xmath77 of @xmath78 \label{fixedq}\ ] ] around @xmath79 ( choosing as normalization @xmath80 ) , @xmath81 several other properties of @xmath77 can be obtained from the fixed point equation ( [ fixedq ] ) or from the recursion ( [ qrecursion ] ) . the simplest one is the limit @xmath82 where @xmath83 is the solution ( @xmath84 ) of @xmath85 this limiting value ( @xmath86 for a population of constant size , i.e. @xmath12 ) is the coefficient of @xmath87 in @xmath3 and so is the fraction of the population whose descendants become extinct : there is a fraction @xmath88 of the population with no children , a fraction of the population with children but no grandchildren , and so on , and the sum of all these contributions gives @xmath83 . equations ( [ qrecursion],[fixedq ] ) have the form of a real space renormalization @xcite . as a consequence , one can predict that for @xmath89 , @xmath90 approaches its limit as a power law , @xmath91 where the exponent @xmath92 must be @xmath93 for the terms of order @xmath94 on both sides of ( [ fixedq ] ) to be equal . for @xmath12 , this gives @xmath95 and ( [ new ] ) implies that at small @xmath1 , the distribution @xmath3 is a power law @xmath96 with @xmath5 given by ( [ betaeq ] ) , in agreement with the results of the simulations shown in figures 1 and 2 . in fact , for @xmath97 , the leading contribution in the difference @xmath98 consistent with ( [ fixedq ] ) is @xmath99 where @xmath100 is an arbitrary periodic function ( not necessarily constant ) of period @xmath101 ( i.e. @xmath102 ) . such periodic amplitudes are often present in the critical behavior of systems which have a discrete scale invariance @xcite . it is easy to calculate numerically the function @xmath77 for all values of @xmath103 from the fixed point equation ( [ fixedq ] ) which relates @xmath103 to points @xmath104 arbitrarily close to @xmath22 , where the linear approximation @xmath105 becomes excellent . using this procedure , we could determine ( figure 3 ) the combination @xmath106 | \lambda|^{-\beta -1}$ ] and the non constant periodic nature of the amplitude @xmath100 is visible if @xmath107 is large enough . the analytic determination of @xmath100 is in principle possible @xcite for @xmath108 close to @xmath101 , but remains difficult for arbitrary @xmath107 . $ ] versus @xmath109 for @xmath110 . we see clearly the periodic nature of the amplitude predicted by ( [ oscillations ] ) . discrepancies at small @xmath111 are due to the fact that the asymptotic regime is not yet reached . at too large @xmath111 , rounding errors in the difference @xmath112 make the resulst noisy and unreliable.,width=340 ] the knowledge of the periodic function @xmath100 determines in principle the whole expansion of @xmath113 in the limit @xmath114 . if we look for a solution of ( [ fixedq ] ) which starts as ( [ oscillations ] ) as @xmath115 , one finds by equating the two sides of ( [ fixedq ] ) order by order in powers of @xmath116 , @xmath117 ^ 3 + \dots \label{expan7}\end{aligned}\ ] ] in addition to the moments ( [ w2lim]-[w4lim ] ) of @xmath3 ( which are given by the expansion ( [ expan ] ) of @xmath77 ) and the exact values ( [ seq],[betaeq ] ) of @xmath83 and @xmath5 , let us just mention two properties of the solution of ( [ fixedq ] ) which we checked by rather involved ways , and that we prefer to leave as conjectures : * @xmath77 is analytic in the whole complex plane of @xmath103 * @xmath77 grows extremely fast ( faster than the exponential of the exponential ... of the exponential of @xmath103 ) as @xmath118 . as a consequence , for large @xmath119 decays faster than any exponential but slower than any stretched exponential ( of exponent larger than 1 ) and even @xmath120 all the discussions of the present section can be repeated in the case of having @xmath121 parents . if we limit ourselves to a population of constant size ( as we did to obtain ( [ wpp ] ) ) , we find that @xmath77 satisfies the same fixed point equation ( [ fixedq ] ) as above with @xmath107 replaced by @xmath121 @xmath122.\ ] ] this means that the distribution of the weights @xmath1 is exactly the same for the cases of ( i ) 2 parents and a population size increasing exponentially by a factor @xmath10 at each generation and ( ii ) a population of constant size with @xmath107 parents per individual . this can be checked by comparing figure 2 and figure 4 , where we show the distributions @xmath20 for a population of constant size @xmath123 and @xmath30 with @xmath124 parents per individual . for a population of constant size with ( a ) @xmath29 and ( b ) @xmath30 when the number @xmath107 of parents is 3 . the generations shown are @xmath125 and @xmath126.,width=340 ] for a population of constant size with ( a ) @xmath29 and ( b ) @xmath30 when the number @xmath107 of parents is 3 . the generations shown are @xmath125 and @xmath126.,width=340 ] despite its simplicity , it is not easy to extract more information on the function @xmath77 and consequently on the distribution @xmath3 from the fixed point equation ( [ fixedq ] ) . there are however two limiting cases around which one can apply a perturbation theory and extract a few more properties of the fixed distribution : @xmath107 close to 1 and @xmath107 very large . one can see from ( [ w2lim]-[w4lim ] ) that when @xmath127 , the successive moments of the weight @xmath1 diverge like @xmath128 . this indicates that if one writes @xmath129 the solution of the fixed point equation ( [ fixedq ] ) can be expanded in the following way @xmath130 where the functions @xmath131 resum the most divergent terms in the perturbative expansion ( [ expan ] ) in the range @xmath132 . if we insert the expansion ( [ expanbis ] ) into ( [ fixedq ] ) we get , by equating the two sides order by order in @xmath133 , a hierarchy of differential equations for the functions @xmath131 which can be solved and lead to @xmath134 \label{f2 } \\ & & f_3 ( y)= { 14 y^3 - 3 y^2 \over 36 \left(1 - { y \over 2 } \right)^3 } + { 17 y^2 - 6 y \over 36 \left(1 - { y \over 2 } \right)^3 } \ln \left [ 1 - { y \over 2 } \right ] \nonumber \\ & & \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; + { y^2 + 2 y \over 36 \left(1 - { y \over 2 } \right)^3 } \ln^2 \left [ 1 - { y \over 2 } \right ] . \label{f3 } \end{aligned}\ ] ] comparing these expressions for large negative @xmath135 with ( [ expan7 ] ) , one gets the expansions of @xmath83 , @xmath5 @xmath136 @xmath137 which both agree with what one would get by directly expanding ( [ seq],[betaeq ] ) . what the small @xmath133 expansion gives us in addition is the function @xmath100 which is found to be a constant function of @xmath138 to all orders in powers of @xmath133 , @xmath139 the non - constant nature of @xmath100 does not show up in the expansion in powers of @xmath133 . it is a non - perturbative contribution ( which vanishes to all orders in @xmath140 ) which could be calculated @xcite using wkb - like techniques @xcite . ( the delta function contribution at @xmath141 is not shown ) for @xmath12 , with @xmath142 and @xmath143 ; @xmath144 , @xmath145 , @xmath146 ; @xmath147 , @xmath148 , @xmath149 ; @xmath150 , @xmath151 , @xmath152 ; @xmath153 , @xmath154 , @xmath152 ; and @xmath155 , @xmath156 , @xmath157 . averages over 1000 realizations have been carried out . the insert shows how the maximum @xmath158 varies with @xmath107.,width=340 ] from ( [ expanbis]-[f2 ] ) and the definition ( [ new ] ) one finds that , for small @xmath133 , the continuous part of @xmath3 is an exponential @xmath159 corrections to this exponential shape are extractable from higher order terms ( @xmath160 , @xmath161 , ) . the other case which can be dealt with perturbatively is the limit of large @xmath107 . if @xmath107 is large and @xmath162 , the solution of ( [ fixedq ] ) is given by @xmath163 \nonumber \\ \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; \ ; + \left[{\lambda^3 \over 2 p^3}+ { \lambda^5 \over 120 p^4 } \right ] + o(p^{-2 } ) \ ; \ ; , \end{aligned}\ ] ] where each term represents a new order in powers of @xmath164 . this implies that @xmath3 can be written in terms of @xmath165 in the range @xmath166 as @xmath167 \ ; \ ; , \label{largep}\end{aligned}\ ] ] where each parenthesis represents a new order in @xmath164 . the gaussian shape in ( [ largep ] ) is not a surprise considering that , for large @xmath107 , each weight becomes the sum of a large number of independent contributions . one property which can be extracted from ( [ largep ] ) is the location of the maximum @xmath158 of @xmath3 @xmath168 figure 5 shows the shapes ( obtained by random samplings populations of constant sizes with @xmath107 parents per individual ) of the distribution @xmath3 for several choices of @xmath107 . the insert shows the values of @xmath158 extracted from these data . they agree with the prediction ( [ wsp ] ) that the maximum approaches @xmath101 with corrections of order @xmath52 as @xmath107 becomes large . we have seen that for simple neutral models of evolution with random mating , the distribution of ancestors repetitions in the genealogical tree of a present individual becomes stationary , with a fixed shape @xmath3 which can be described by a fixed point equation of the type ( [ fixedq ] ) . this shape is the same if one considers a population increasing exponentially at rate @xmath10 per generation with two parents per individual or a population of constant size with @xmath107 parents per individual . the fixed point equation ( [ fixedq ] ) allows one to determine exactly the exponent @xmath5 which characterizes @xmath3 at small @xmath1 . the determination of @xmath5 from ( [ fixedq ] ) is very reminiscent of the way one finds exponents in the renormalization group approach of critical phenomena . other properties ( large @xmath1 behavior , amplitude of the power law , ) of the fixed distribution @xmath3 are in principle extractable from ( [ fixedq ] ) but are more difficult to obtain than the exponent @xmath5 . the present work admits several extensions . in particular , one may consider the case where the probabilities @xmath64 ( that an individual has @xmath63 children ) is arbitrary ( instead of poissonian as in ( [ poisson ] ) ) . the fixed point equation ( [ fixedq ] ) becomes then simply @xmath169 and starting from this new fixed point equation , one can essentially repeat all the above calculations , including the determination of the exponent @xmath5 . if all the @xmath64 vanish for @xmath170 , one can see that for large @xmath103 , @xmath171 consequently , the distribution @xmath3 becomes a stretched exponential for large @xmath1 , @xmath172 recursions similar to ( [ wp ] ) describe the distribution of constraints in granular media @xcite . in such cases , the number of grains in direct contact and supporting the weight of a given grain is variable . this would correspond to considering that the number @xmath121 of parents is no longer constant over the whole population but may vary from individual to individual . finally let us mention that an interesting aspect of the problem is the calculation of the correlations between the genealogies of several contemporary individuals . one can show @xcite that for large @xmath0 , the weights of all the ancestors of two distinct individuals in the same population become the same after a number of generations @xmath173 . in this appendix we show , by calculating moments of the weights @xmath174 , that correlations become negligible in the r.h.s . of ( [ wp ] ) and ( [ wpp ] ) . it is convenient to rewrite ( [ wp ] ) as @xmath176 where @xmath177 for the random parent model of section 2 ( where each parent of @xmath178 is chosen at random among all the individuals of the previous generation ) , @xmath179 with probability @xmath180 , @xmath181 with probability @xmath182 and @xmath183 with probability @xmath184 ( as we did not exclude choosing the same parent twice ) . moreover there is no correlation between @xmath185 and @xmath186 if @xmath187 . lastly @xmath185 and @xmath188 are correlated for @xmath189 and @xmath190 this correlation together with @xmath191 when used in ( [ wp1 ] ) leads to @xmath192 as expected , since the definition ( [ wdef ] ) of @xmath1 was chosen to keep @xmath193 , and @xmath194 where @xmath195 ( the index @xmath196 has been omitted for simplicity ) . from ( [ initial],[evolution],[wdef ] ) , we know that @xmath197 , and @xmath198 . thus for @xmath195 @xmath199 and ( [ w2a ] ) becomes @xmath200 so far this evolution equation is exact . if we consider that all the @xmath6 s are very large , ( [ w2b ] ) becomes @xmath201 so that for large @xmath0 ( in fact @xmath0 should not be too large to keep @xmath6 large enough , more precisely @xmath0 should be such that @xmath202 ) , the second moment of @xmath1 has a limiting value @xmath203 and we see from ( [ wwb ] ) that @xmath204 when one repeats the above calculation for higher correlations ( we did it up to three - point correlations ) , one finds that the correlations between the terms in the r.h.s . of ( [ wp1 ] ) are negligible . this indicates that these correlations can be neglected ( of course a complete proof that all correlations are negligible in the scaling limit would be much better than our guess based on the computation of the lowest correlations ) . one can repeat the above calculation of correlations for several variants of the model , like those discussed at the end of section 2 . the exact formulae ( [ w2a],[wwa],[w2b ] ) are modified but one always find that , in the scaling regime , they reduce to ( [ w2c],[wwe ] ) , meaning that the correlations could be ignored . let us consider only the case where each individual has @xmath121 parents . to keep the notations simple , we will limit the calculation to the case of a population of constant size @xmath205 one can then follow the same steps as above . starting from ( [ wpp ] ) , one replaces ( [ wp1 ] ) by @xmath206 the correlations ( [ cor1]-[cor4 ] ) become in this case @xmath207 and ( [ w2a],[wwa ] ) read @xmath208 for large @xmath0 and large @xmath34 , we see ( using the fact that @xmath209 ) that @xmath210 and @xmath211 as @xmath212 . this again indicates that correlations can be neglected for large @xmath0 and large @xmath34 . interesting discussions with ugo bastolla and vincent hakim are gratefully acknowledged . scm acknowledges support from the alexander von humboldt foundation ( germany ) . c. domb , _ the critical point_. taylor and francis , london 1996 . mandelbrot , _ the fractal geometry of nature_. freeman , new york 1983 . m. feigenbaum , j. stat . * 21 * ( 1979 ) 669 - 706 ; p. cvitanovi , _ universality in chaos _ adam hilger , 1984 ( and references therein ) . b. duplantier , physica a * 163 * ( 1990 ) 158 - 182 ( and references therein ) . p. bak , c. tang , k. wiesenfeld , phys . * 59 * ( 1987 ) 381 - 384 ; p. bak , _ how nature works _ oxford university press , 1997 ; d. dhar , _ studying self - organized criticality with exactly solved models _ , preprint cond - mat/9909009 . b. derrida , physica d * 103 * ( 1997 ) 466 - 477 ( and references therein ) . b. derrida , s.c . manrubia , and d.h . zanette , phys . rev . lett . * 82 * ( 1999 ) 1987 - 1990 . m. serva and l. peliti , j. phys . a : math . gen . * 24 * ( 1991 ) l705-l709 . t.e . harris , _ the theory of branching processes_. springer - verlag ohg , 1971 . burkhardt and j.m.j . van leeuwen , _ real space renormalization _ , vol . 30 . springer , berlin 1982 . b. derrida , a. gervois , and y. pomeau , j. phys . a * 12 * ( 1979 ) 269 - 296 ; b. derrida , c. itzykson , and j.m . luck , com . * 94 * ( 1984 ) 115 - 132 ; d. sornette , phys . rep . * 297 * ( 1998 ) 239 - 270 , ( and references therein ) . v. hakim , private communication v. hakim and k. mallick , nonlinearity * 6 * ( 1993 ) 57 - 70 . coppersmith , c .- h . liu , s.n . majumdar , o. narayan , and t.a . witten , phys . e * 53 * ( 1996 ) 4673 - 4685 . b. derrida , s.c . manrubia , and d.h . zanette , _ on the genealogy of a population of biparental individuals _ , preprint 1999 .

we calculate the probability distribution of repetitions of ancestors in a genealogical tree for simple neutral models of a closed population with sexual reproduction and non - overlapping generations . each ancestor at generation @xmath0 in the past has a weight @xmath1 which is ( up to a normalization ) the number of times this ancestor appears in the genealogical tree of an individual at present . the distribution @xmath2 of these weights reaches a stationary shape @xmath3 , for large @xmath0 , i.e. for a large number of generations back in the past . for small @xmath1 , @xmath3 is a power law ( @xmath4 ) , with a non - trivial exponent @xmath5 which can be computed exactly using a standard procedure of the renormalization group approach . some extensions of the model are discussed and the effect of these variants on the shape of @xmath3 are analysed . genealogy , critical phenomena , renormalization group .

the work of q. shafi was supported in part by the us department of energy grant no . de - fg02 - 91er40626 . h. georgi , c. jarlskog , phys . b 88 ( 1979 ) 279 ; j. harvey , p. ramond , d. reiss , phys . lett . b 92 ( 1980 ) 309 ; s. dimopoulos , l.j . hall , s. raby , phys . 68 ( 1992 ) 1984 ; v. barger et . al . , phys . 68 ( 1992 ) 3394 ; h. arason et . al . , phys . d 47 ( 1993 ) 232 . s. dimopoulos , f. wilczek , in erice summer lectures , plenum , new york , 1981 ; h. georgi , phys . b108 ( 1982 ) 283 ; b. grinstein , nucl . b206 ( 1982 ) 387 ; a. masiero el al . b115 ( 1982 ) 380 ; j.l . lopez , d.v . nanopoulos , phys . d53 ( 1996 ) 2670 . j. chkareuli , a. kobakhidze , phys . b 407 ( 1997 ) 234 ; j. chkareuli , i. gogoladze , a. kobakhidze , phys . rev . 80 ( 1988 ) 912 ; j. chkareuli et al . d 62 ( 2000 ) 015014 ; nucl . b 594 ( 2001 ) 23 . r. barbieri , g. dvali , a. strumia , nucl . b391 ( 1993 ) 487 ; r. barbieri , g. dvali , m. moretti , phys . b312 ( 1993 ) 137 ; r. barbieri et al . , nucl.phys . b432 ( 1994 ) 49 ; z. berezhiani , phys . b355 ( 1995 ) 481 ; z. berezhiani , c. csaki , l. randall , nucl . b444 ( 1995 ) 61 ; g. dvali , s. pokorski , phys . 78 ( 1997 ) 807 ; q. shafi , z. tavartkiladze , nucl . b 573 ( 2000 ) 40 . g. dvali , phys . b 324 ( 1994 ) 59 ; i. gogoladze , a. kobakhidze , z. tavartkiladze , phys . lett . b 372 ( 1996 ) 246 ; a. kobakhidze , phys . lett . b 391 ( 1997 ) 335 ; z. tavartkiladze , phys . lett . b 392 ( 1997 ) 360 .

we present an example of susy @xmath0 gut , which predicts an excellent value for @xmath1 , in comparison with the value @xmath2 of minimal susy @xmath3 . a crucial role is played by the vectorlike multiplets from the matter sector , whose masses lie below the gut scale . for a realistic pattern of fermion masses , the adjoint scalar of @xmath0 has vev along the @xmath4 direction . this also offers a natural resolution of the doublet - triplet ( dt ) splitting through the pseudo goldstone boson mechanism . epsf -2 cm -1.5 cm ( ba-01 - 21 + * realistic supersymmetric @xmath0 * qaisar shafi@xmath5 and zurab tavartkiladze@xmath6 @xmath7_bartol research institute , university of delaware , newark , de 19716 , usa + @xmath8 institute of physics , georgian academy of sciences , 380077 tbilisi , georgia _ + the minimal susy @xmath3 model suffers from a variety of nagging problems . for instance , the measured value of the strong coupling @xmath9 @xcite , while the predicted value is @xmath10 @xcite . it also predicts the wrong asymptotic relations @xmath11 , @xmath12 . and finally , although susy guarantees stability of scales against radiative corrections , the origin of dt splitting remains unexplained . in attempting to resolve these problems , one can either consider some extended versions of @xmath3 or an alternative gut scenario . in fact , for obtaining a desirable value of @xmath13 , some additional states below the gut scale could play an important role @xcite . for realistic fermion masses , either a scalar @xmath14 plet @xcite or additional fermionic states @xcite can be introduced . within @xmath3 , solution of the dt splitting problem requires a rather complicated @xmath15 set of scalars , which turn out to be crucial for realization of the missing partner mechanism @xcite . replacing @xmath3 with @xmath16 , one can achieve dt splitting through the missing vev mechanism @xcite . ] . very attractive and promising scenarios are those in which the light higgs doublets emerge as pseudo - goldstone bosons ( pgb ) . this idea is easily realized within @xmath0 @xcite , @xmath17 @xcite or flipped @xmath0 @xcite models . also , @xmath0 scenarios with additional custodial symmetries can provide a natural understanding of dt splitting @xcite . in this letter we show how these three problems could be simultaneously resolved by considering an @xmath0 gut . the value of @xmath13 , it turns out , is closely tied with the matter sector , and is expressed through some asymptotic mass relations . it is interesting to note that a realistic pattern of fermion masses unequivocally requires the vev of the adjoint higgs to be along the @xmath4 direction . this also permits realization of the pgb mechanism @xcite for achieving a natural dt splitting . consider the susy @xmath0 gut with chiral ` matter ' multiplets @xmath18 per generation . in terms of @xmath3 : @xmath19 , @xmath20 ( and same for @xmath21 ) . thus , we have the additional @xmath3 @xmath22 vectorlike states , which decouple after @xmath0 breaking . at first glance , since they are complete @xmath3 plets , one may think that the picture of gauge coupling unification will not be altered at one loop level . however , it turns out that the doublet and triplet fragments from these additional @xmath23 ( @xmath24 ) plets are split in mass . this happens because , in order to get a realistic pattern of down quark and charged lepton masses , we somehow must remove the degeneracy between their mass matrices . if this is done , then the heavy vectorlike doublet and triplet states also will acquire different masses , and their ratios will be expressed through asymptotic mass relations of down quarks and charged leptons , giving rise to the possibility of predicting @xmath13 . the relevant @xmath0 invariant couplings , in lowest order , are of the form @xmath25 , where @xmath26 ( @xmath27 ) is an antisextet ( sextet ) scalar field . in order to avoid the wrong asymptotic relations @xmath11 , @xmath12 we will insert in these couplings the @xmath0 adjoint scalar @xmath28 [ this can be realized through a @xmath29 symmetry @xmath30 , @xmath31 . for a transparent demonstration , let us first consider the case of one generation . the relevant couplings are : 15_ij^i_m |6^m|h^j+|6^j|h^m+ |6^m|h^j+ |6^j|h^m , where @xmath32 are @xmath0 indices , @xmath33 - are dimensionless couplings , and @xmath34 is some cutoff mass scale . @xmath35 and @xmath26 have vevs of the same order ( @xmath36 ) , and the light higgs doublet @xmath37 is suppressed by equal weights in these plets . it is easy to verify that the relevant terms are built with the higgs doublet @xmath37 extracted from @xmath26 , and we will ignore terms in which the doublets from @xmath35 participate ( such terms do not lead to light fermion masses to be identified as quarks and leptons , will couple with decoupled states ) . from ( [ yuk ] ) , we have : @xmath38 it is obvious that the symmetry breaking patterns $ su(5)\times u(1)$ and $ su(3)\times su(3)\times u(1)$ are not plausible , since , in these cases , we either have degeneracy between $ \hat{m}_d$ and $ \hat{m}_e$ or the determinants in ( \ref{dets } ) are zero [ in the latter case some quark and lepton states are massless ] . we therefore conclude that the only possible $ \lan \si \ran $ vev which can lead to a realistic fermion mass pattern is $ su(4)\times su(2)\times u(1)($e.g . $ \si_c=\si_6=-\si_w/2 $ ) \beq \lan \si \ran = { \rm diag}(1 , 1 , 1 , -2 , -2 , 1)\cdot v~. \la{421dir } \eeq using ( \ref{421dir } ) in ( \ref{dets } ) , one obtains \beq { \rm det } ( \hat{m}_e ) = -2\cdot{\rm det } ( \hat{m}_d ) ~ , \la{reldet } \eeq which implies $ m_e^{(0)}m^l=2m_d^{(0)}m^{d^c}$ , where $ m^l$ , $ m^{d^c}$ are the masses of the heavy doublet and triplet components respectively , and $ m_e^{(0)}$ , $ m_d^{(0)}$ denote the asymptotic values of charged lepton and down quark mases . therefore , \beq \fr{m^l}{m^{d^c}}=2\fr{m_d^{(0)}}{m_e^{(0)}}~. \la{relmass } \eeq knowing the asymptotic value of $ \fr{m_d^{(0)}}{m_e^{(0)}}$ for a given generation , we calculate through ( \ref{relmass } ) the ratio $ m^l / m^{d^c}$. the latter give us possibility to predict the value of $ \al_s(m_z)$. analagous results can be obtained for the case with three generations , and as we will see , even inclusion of intergeneration mixings do not modify the picture . % couplings $ \al , \dots \bt\hs{0.4mm}'$ % in ( \ref{yuk } ) will be $ 3\times 3 $ matrices and instead of ( \ref{matd } ) , ( \ref{matl } ) we will have $ 6\times 6 $ matrices . using ( \ref{421dir } ) , the appropriate mass matrices are : % \left(\frac{x}{m_p}\right)^{r ' } it is not difficult to find a relation between the determinants of matrices in ( \ref{bigmats } ) . recall that determinants remain unchanged by making some linear manipulations with their rows and coulomns . more precisely : % \left(\frac{x}{m_p}\right)^{r ' } \beq { \rm det } ( \hat{\cal m}_e ) = -8\cdot{\rm det } ( \hat{\cal m}_d ) ~. \la{relbigdets } \eeq therefore , $ m_e^{(0)}m_{\mu } ^{(0)}m_{\tau } ^{(0)}m^l_1m^l_2m^l_3= 8m_d^{(0)}m_s^{(0)}m_b^{(0)}m^{d^c}_1m^{d^c}_2m^{d^c}_3 $ , where $ m^l_i$ , $ m^{d^c}_i$ denote the masses of heavy doublets and triplets of the corresponding generation . finally : \beq \fr{m^l_1m^l_2m^l_3}{m^{d^c}_1m^{d^c}_2m^{d^c}_3}= 8\l \fr{m_dm_sm_b}{m_em_{\mu } m_{\tau } } \r^{(0)}~. \la{rel3mass } \eeq we will see below that the value of $ \al^{-1}_s(m_z)$ will depend logarithmically on the ratio in ( \ref{rel3mass } ) . the solutions of the rges are \cite{alpsu5 } : \beq \al^{-1}_g=\al^{-1}_a-\fr{b_a}{2\pi } \ln \fr{m_g}{m_z}- \fr{b^l_a}{2\pi } \si_i \ln \fr{m_g}{m^l_i}- \fr{b^{d^c}_a}{2\pi } \si_i \ln \fr{m_g}{m^{d^c}_i}+ \de_a+\de_a~ , \la{rge } \eeq where $ \al_g$ is the gauge coupling at the gut scale , $ \al_a$ the gauge coupling at $ m_z$ ( $ \al_{1 , 2 , 3}$ are gauge couplings of $ u(1)$ , $ su(2)_w$ and $ su(3)_c$ respectively ) , while \beq b_a=(\fr{33}{5 } , 1 , -3)~,~~ b^l_a=(\fr{3}{5 } , 1 , 0)~,~~ b^{d^c}_a=(\fr{2}{5 } , 0 , 1)~. \la{bfacs } \eeq the $ \de_a$ include all possible threshold corrections and two loop effects of mssm . $ \de_a$ denote the difference between mssm and the present model of the gauge coupling running from $ m_1^{d^c}$ ( lowest possible intermediate scale ) up to $ m_g$ in two loop approximation , \beq \de_a=\fr{1}{4\pi } \l \fr{b_{ab}^{\rho } } { b_b^{\rho } } \ln \fr{\al_b(m_{\rho } ) } { \al_b(m_{\rho+1})}-\fr{b_{ab}}{b_b } \ln\fr{\al_b(m_1^{d^c})}{\al_g^0 } \r ~ , \la{twoloop } \eeq where summation over $ \rho $ and $ b$ indices is implied . $ \rho $ enumerates the heavy vectorlike doublet and triplet states below the gut scale , and $ m_{\rho } $ and $ b^{\rho } _ a$ , $ b^{\rho } _ { ab}$ are the corresponding mass scale and b - factors ( which depend on energy scale ) respectively . $ b_{ab}$ denote two loop b - factors of mssm . in ( \ref{twoloop } ) the appropriate couplings are calculated in one loop approximation . $ \al_g^0 $ is the gauge coupling of mssm at $ m_g$. for the time being in ( \ref{rge } ) we will ignore $ \de_a$. calculating the combination $ 12\al^{-1}_2 - 5\al^{-1}_1 - 7\al^{-1}_3 $ and taking into account ( \ref{rel3mass } ) , one obtains : \beq \l \al^{-1}_s\r'=\l \al^{-1}_s\r^0+\fr{9}{14\pi}\ln \fr{m^l_1m^l_2m^l_3}{m^{d^c}_1m^{d^c}_2m^{d^c}_3}=\l \al^{-1}_s\r^0 + \fr{9}{14\pi}\ln \l 8\fr{m_dm_sm_b}{m_em_{\mu } m_{\tau } } \r^{(0)}~ , \la{alp3 } \eeq where $ \l \al^{-1}_s\r^0=\fr{1}{7}\l 12\al^{-1}_2 - 5\al^{-1}_1 + 12\de_2 - 5\de_1 - 7\de_3 \r $ corresponds to the value of $ \al_s$ obtained for mssm ( or mssu5 ) . the prime on $ \al_s$ indicate that it is calculated ignoring two loop effects coming from $ \de_a$ terms . employing the reasonable asymptotic relations \beq \al_s^{-1}=\l \al_s^{-1}\r'+\de~ , \la{als2loop } \eeq where $ \de=\fr{1}{7}(12\de_2 - 5\de_1 - 7\de_3)$. in order to calculate $ \de_a$ in ( \ref{twoloop } ) , we have to know the masses of doublet and triplet vectorlike states . from ( \ref{matd } ) , ( \ref{matl } ) and ( \ref{relbigdets } ) , it is natural to assume that for each family we have $ m_i^{d^c}\simeq m_g\lam_d^i$. also for each family we will assume relation ( \ref{relmass } ) which , taking into account ( \ref{asym } ) , gives $ m_3^l=2m_3^{d^c}$ , $ m_2^l=2m_2^{d^c}/3 $ , $ m_1^l=6m_1^{d^c}$. recall that for the pgb $ su(6)$ scenario , the preferred value of $ \tan \bt$ is order unity \cite{pgbsu62}-\cite{pgbsu3 } , so that $ \lam_b\sim \lam_{\tau } \sim 10^{-2}$. we also have the measured hierarchies between down quark yukawa couplings : namely , $ \lam_d : \lam_s : \lam_b \sim \ep^5:\ep^2:1 $ , where $ \ep\simeq 0.2$. taking all this into account , for the mass spectra of the vectorlike states , it is quite natural to have : \begin{equation } \begin{array}{cc } b_{ab } \hs{-1mm}=\!\!\!\!\hs{-1 mm } & \left ( \begin{array}{ccc } \hs{-1 mm } \fr{199}{25}~ , & \fr{27}{5}~ , & \fr{88}{5 } \\ \hs{-1 mm } \fr{9}{5}~ , & 25~,&24 \\ \hs{-1mm}\fr{11}{5}~ , & 9~,&14 \end{array}\hs{-1.5mm}\right)\hs{0.1 mm } , \end{array } \begin{array}{cc } b_{ab}^{d^c}\hs{-1mm}=\!\!\!\!\!\hs{-1 mm } & { \left(\begin{array}{ccc } \hs{-1 mm } \fr{8}{75}~ , & 0~,&\fr{32}{15 } \\ \hs{-1 mm } 0~ , & 0~,&0 \\ \hs{-1 mm } \fr{4}{15}~ , & 0~,&\fr{34}{3 } \end{array}\hs{-1.5mm}\right)\hs{0.1 mm } , } \end{array } % \begin{array}{cc } % { \cal f}^{\hs{1mm}c}\hs{-1mm}=\!\!\!\!\hs{-1 mm } & \left % ( \begin{array}{ccc } % \hs{-1mm}\fr{9}{25}~ , & \fr{9}{5}~ , & 0 % \\ % \hs{-1 mm } \fr{3}{5}~ , & 7~ , & 0 % \\ % \hs{-1 mm } u_3^{\hs{1mm}c}~ , & d_3^{\hs{1mm}c } % \\ % \hs{-1 mm } \nu^{\hs{1mm}c}~ , & e^{\hs{1mm}c } % \end{array}\hs{-1.5mm}\right)\hs{0.1 mm } , % \end{array } \begin{array}{cc } b_{ab}^l\hs{-1mm}=\!\!\!\!\!\hs{-1 mm } & { \left(\begin{array}{ccc } \hs{-1 mm } \fr{9}{25}~ , & \fr{9}{5}~ , & 0 \\ \hs{-1 mm } \fr{3}{5}~ , & 7~ , & 0 \\ \hs{-1 mm } 0~ , & 0~ , & 0 \end{array}\hs{-1.5mm}\right)\hs{0.1mm}. } \end{array } \label{bb}\ ] ] from ( [ twoloop ] ) , taking into account ( [ spectra])-([bb ] ) , we obtain @xmath39 , and according to ( [ als2loop ] ) @xmath40 in excellent agreement with the experimental data @xcite . numerical calculations confirm these estimations . the unification picture of gauge couplings is presented on fig . 1 . as far as the up - type quark sector is concerned , the relevant couplings for their mass generation are @xmath41 , where @xmath42 are family indices and @xmath34 is some cutoff mass scale of the order of @xmath43 . these operators could emerge through exchange of some additional states with masses @xmath44 @xcite . as we have demonstrated , a realistic fermion mass pattern is realized when @xmath45 aligns along the @xmath4 direction ( [ 421dir ] ) . this indeed also happens to be the vev direction required for realization of the pgb mechanism within @xmath0.we refer the reader to @xcite , where detailed studies of this question are presented . in conclusion , let us note that in the present scenario , it is possible to invoke flavor symmetries , the simplest being @xmath46 for a natural understanding of hierarchies between charged fermion masses and the ckm matrix elements . the various neutrino oscillation scenarios are considered in @xcite . if the flavor @xmath46 turns out to be anomalous , it also helps in achieving an ` all order ' dt hierarchy ( see last two refs . in @xcite ) .

our universe has been evolving for 13.8gyr . over these years many stars formed and ended their lives enriching the interstellar medium ( ism ) , and in consequence enriching the universe ( ( * ? ? ? * planck collaboration et al . very few elements have been around since the beginning , formed by the big bang nucleosynthesis ( bbn ) . bbn is responsible for the formation of most of the helium isotope ( @xmath2he ) in the universe , along with small amounts of deuterium ( d ) , the helium isotope ( @xmath0he ) , and a very small amount of the lithium isotope ( @xmath3li ) ( ( * ? ? ? * alpher , r. a. et al . 1948 ) ) . the predicted abundance ( by number ) of @xmath0he ( relative to h ) formed by the bbn ( just after a few minutes after the big bang ) is 1.0@xmath410@xmath5 ( ( * ? ? ? * karakas , a. i. & lattanzio , j. c. 2014 ) ) . this abundance depends only on the parameter of the current density of baryonic matter . the present interstellar @xmath0he abundance , as per all the light elements , comes from a combination of bbn and stellar nucleosynthesis ( ( * ? ? ? * wilson , t. l. & rood , r. 1994 ) ) . hii regions are young objects compared with the age of the universe , and represent zero - age objects . their @xmath0he abundance is the result of 13.8gyr of galactic chemical evolution . in between is the solar system , which traces abundances at the time of its formation , 4.6gyr ago . observed values in pre - solar material ( ( * ? ? ? * geiss , j. 1993 ) ) and the ism ( ( * ? ? ? * gloeckler , g. & geiss , j. 1996 ) ) imply that @xmath0he / h@xmath6([email protected])@xmath410@xmath5 . these values from the ism and pre - solar material are approximately twice of the bbn , implying that the @xmath0he abundance has increased a little in the last 13.8gyr . on the other hand any hydrogen - burning zone of a star which is not too hot ( @xmath87@xmath410@xmath9k ) will produce @xmath0he via the _ p - p _ chain , implying that stars with masses @xmath102.5m@xmath11 are net producers of @xmath0he . for these stars , _ p - p _ burning is rapid enough to produce d _ in situ _ , and enable the production of @xmath0he ( d+ p @xmath12 @xmath0he + @xmath13 ) . stellar evolution models indeed predict the formation of @xmath0he in significant amounts by stars of 12.5m@xmath11 , with an abundance of @xmath0he / h @xmath14 ( ( * ? ? ? * bania , t. m. et al . 2010 ) ) , which would have raised the current @xmath0he abundance to @xmath0he / h@xmath15 , substantially higher than observed ( ( * ? ? ? * karakas , a. i. & lattanzio , j. c. 2014 ) ) . @xcite presented the @xmath0he problem " . according to standard models of stellar nucleosynthesis there should be a @xmath0he / h abundance gradient in the galactic disk and the proto - solar @xmath0he / h value should be less than what is found in the present ism . observations of the @xmath0he abundance in hii regions show almost no enrichment above the bbn value ( ( * ? ? ? * ; * ? ? ? * rood , r. t. et al . 1979 ; bania , t. m. et al . 2007 ) ) . for the @xmath0he problem to be solved , the vast majority of low - mass stars should fail to enrich the ism . one suggestion to solve this problem is by adding extra mixing in the red giant branch ( rgb ) stage . this extra mixing adds to the standard first dredge - up to modify the surface abundances . @xcite estimated that while 90% of the @xmath0he is destroyed in 1m@xmath11 stars , only 40 - 60% is destroyed in a 2m@xmath11 star model , depending on the speed of mixing . the abundance of @xmath0he can only be derived from the hyperfine transition at the rest frequency of 8.665ghz . detecting @xmath0he@xmath1 in pne challenges the sensitivity limits of all existing radio telescopes . composite @xmath0he@xmath1 average spectrum for six planetary nebulae ( ngc 3242 , ngc 6543 , ngc 6720 , ngc 7009 , ngc 7662 , and ic 289 ) , using effelsberg , arecibo and the green bank telescope observations . they all consistently show @xmath0he@xmath1 emission at the @xmath161mk level . using the same telescopes later on a detection was obtained in ngc 3242 and j320 . these detections and the composite spectrum measurements translates into an abundance of @xmath0he / h of a few 10@xmath17 ( ( * ? ? ? * ; * ? ? ? * ; * ? ? ? * balser , d. s. et al . 1997 , 1999 , 2006 ) ) . later on more effort was put into this problem , using the very large array , for several tens of hours 3 planetary nebulae were observed ( ic418 , ngc 6572 , and ngc 7009 ) , unfortunately no line was detected and only upper limits in the abundance were obtained ( ( * ? ? ? * guzman - ramirez , l. et al . 2013 ) ) . only just recently ( beginning of this year ) @xcite using the deep space station 63 antenna from the madrid deep space communications complex , another @xmath0he@xmath1 detection was observed in the planetary nebula ic418 . with more than 20 years of looking for these detections , so far we have only 3 objects with more than a 3@xmath18 detection of @xmath0he : j320 , ngc 3242 , and ic 418 . he@xmath1 line detected in planetary nebulae . the sources are ngc3242 ( balser et al . 1999 ) , j320 ( balser et al . 2006 ) , and ic418 ( this work ) . in order to facilitate the comparison , the abscissa is presented as radial velocities relative the lsr systemic velocity of each source ( dashed - line ) , and the intensity scale as fractions of the peak intensity of each source.,width=283,height=321 ] two aspects of these detections needs to be considered ; firstly , the derived @xmath0he@xmath1 abundance is well above model expectations for all the three objects . this tell us that the stars do produce and release its @xmath0he into the interstellar medium . secondly , the @xmath0he@xmath1 line profile for all the three objects shows a double peaked shape . although the full width of the @xmath0he@xmath1 line is consistent with the optical expansion velocity , the profile differs from that of the recombination lines , peaking at the outermost velocities . the two other planetary nebulae have reported @xmath0he@xmath1 detections : j320 ( ( * ? ? ? * balser , d. s. et al . 2006 ) ) and ngc 3242 ( ( * ? ? ? * ; * ? ? ? * balser , d. s. et al . 1997 , 1999 ) ) . both have double peaked profiles , similar to ic418 ( fig . [ profiles ] ) . both objects have haloes , in the case of ngc3242 possibly as large as 18 by 24arcmin diameter . whether helium in such a halo could be photo - ionized by the star is not clear . a double - peaked profile could arise from an expanding detached shell , which is larger than the beam of the telescope . in this case , the outer components of the profile , which differ from the recombination lines , come from this large region , whilst the central part of the profile arises from the inner , ionized nebula . @xcite also proposed a contribution from a large , low density halo . the emission in the @xmath0he@xmath1 hyperfine line scales with @xmath19 ( column density ) , and that of recombination line with @xmath20 . therefore , a low density but high mass halo could explain the difference in profiles . the contribution of pne to the @xmath0he abundance is crucial for understanding the galactic chemical evolution . the number density of @xmath0he@xmath1 atoms , @xmath21(@xmath0he@xmath1 ) , can be obtained by dividing the column density @xmath22(@xmath0he@xmath1 ) by the averaged optical path , @xmath23 , through the source . representing a pn as a homogeneous sphere of radius @xmath24 , where @xmath25 is the distance , the optical path at angle from the centre @xmath26 is @xmath27 , and the optical path averaged over the source becomes @xmath28 . to calculate the fractional @xmath0he abundance we divide by the h@xmath1 density . figure [ abundances ] shows the @xmath0he abundances of the pn ic418 ( both purple crosses , taken from @xcite ) . for comparison the upper limits calculated from @xcite observations are presented using the red arrows . @xcite observations are shown in green : the cross represents j320 and the arrows are the upper limits , where the mass estimates are from @xcite . the stellar evolution models are also presented in the same figure . he ( relative to h ) for the pn ic418 ( both purple crosses , for upper and lower estimate ) , the size of the bars correspond to the uncertainties . red arrows are the upper limits from @xcite . the green cross represents j320 and the green arrows represent the sample of 6 pne from @xcite . the curves for pop i ( light - blue line ) and ii ( red dashed - line ) show the standard abundance of @xmath0he taken from @xcite . the purple dashed - line labelled extra mixing represents the results of models including deep mixing by @xcite , and the blue line shows the stellar models using thermohaline mixing ( ( * ? ? ? * charbonnel , c. & zahn , j .- the black line is the primordial value of @xmath0he / h from the bbn ( ( * ? ? ? * karakas , a. i. & lattanzio , j. c. 2014)).,width=283,height=302 ] the values observed in these objects proves that @xmath0he is produced at the centre of low - mass stars , and is ejected into the interstellar medium at the end of its life . however , the large amounts of @xmath0he produced in these models is at odds with the abundance of @xmath0he observed in the interstellar medium and the solar system . further research on other planetary nebulae will be needed to solve the @xmath0he problem , but as the emission is extremely weak and hard to detect , many observing hours are needed , or telescopes with higher sensitivity like the square kilometre array will be needed to do follow this research .

the @xmath0he isotope is important to many fields of astrophysics , including stellar evolution , chemical evolution , and cosmology . the isotope is produced in stars which evolve through the planetary nebula phase . planetary nebulae are the final evolutionary phase of low- and intermediate - mass stars , where the extensive mass lost by the star on the asymptotic giant branch is ionised by the emerging white dwarf . this ejecta quickly disperses and merges with the surrounding ism . the abundance of @xmath0he can only be derived from the hyperfine transition of the ionised @xmath0he , which is represented as @xmath0he@xmath1 , these transition can be observed in the radio at the rest frequency of 8.665ghz . @xmath0he abundances in pne can help test models of the chemical evolution of the galaxy . many hours have been put into trying to detect this line , using telescopes like effelsberg a 100 m dish from the max planck institute for radio astronomy , the national radio astronomy observatory ( nrao ) 140-foot telescope , the nrao very large array , the arecibo antenna , the green bank telescope , and only just recently , the deep space station 63 antenna from the madrid deep space communications complex .

in the present paper the simplest three - dimensional system , describing the convection of fluid within an ellipsoidal rotating cavity , is considered . this system , suggested by glukhovsky and dolghansky @xcite , is as follows @xmath0 where @xmath1 , @xmath2 , @xmath3 are positive numbers . after the change of variables @xmath4 system takes the form @xmath5 where @xmath6 this system is a generalization of a classical lorenz system @xcite . system with the parameters @xmath7 , @xmath1 , @xmath8 is mentioned first in @xcite and after the corresponding change of variables @xcite can be transformed to the rabinovich system of waves interaction in plasma @xcite . consider system under the assumption that @xmath7 , @xmath1 , @xmath9 , @xmath10 are positive . in this case by @xcite one obtains the following : if @xmath11 , then has a unique equilibrium @xmath12 ( the trivial case ) . if @xmath13 , then has three equilibria : @xmath12 and @xmath14 , where @xmath15 and the real number @xmath16 is defined as @xmath17.\ ] ] consider a differential equation @xmath18 with a continuously differentiable vector function @xmath19 . assume that for any initial value @xmath20 there exists a unique solution of @xmath21 defined for @xmath22 . here let @xmath24 be an invariant set , i.e. @xmath25 for all @xmath22 . suppose @xmath26 is the jacobian matrix of @xmath19 @xmath27 consider a nonsingular @xmath28-matrix @xmath29 . suppose that @xmath30 are eigenvalues of the matrix @xmath31 here * denotes a matrix transposition . * @xcite*[theorem : th1 ] given an integer @xmath32 $ ] and @xmath33 there is a continuously differentiable scalar function @xmath34 and a nonsingular matrix @xmath29 such that @xmath35 in this case the lyapunov dimension of the compact set @xmath24 is estimated as follows @xmath36 here @xmath37 is a derivative of the function @xmath38 with respect to the vector field @xmath39 : @xmath40 * @xcite*[theorem : th2 ] suppose that there is a continuously differentiable scalar function @xmath38 and a nonsingular matrix @xmath29 such that @xmath41 then any solution of system , bounded on @xmath42 tends to an equilibrium as @xmath43 . thus , if condition holds , then an attractor of system coincides with its stationary set . in @xcite it is proved that system is dissipative , i.e. it possesses a bounded absorbing set and , thus , has an attractor . by theorems [ theorem : th1 ] and [ theorem : th2 ] it can be formulated the assertion , concerning the lyapunov dimension of an attractor ( denote it by @xmath24 ) of system . this assertion is a generalization of the result obtained in @xcite for @xmath44 . [ theorem : th3 ] suppose that either the inequality @xmath45 or the inequalities @xmath46 , @xmath47 are valid . + if @xmath48 then any solution of system , bounded on @xmath49 , tends to an equilibrium as @xmath43 . + if @xmath50 then @xmath51 we use the matrix @xmath52 then the eigenvalues of the corresponding matrix are the following @xmath53^{\frac{1}{2}}. & \end{aligned}\ ] ] the goal is to obtain a lyapunov - like function that allows one to check property of theorem [ theorem : th1 ] and property of theorem [ theorem : th2 ] . for considered eigenvalues one can obtain the following relation @xmath54^\frac{1}{2 } + \\ & & + \frac{2(1-s)}{\left[(\sigma - 1)^2 + a\left(\frac{\sigma}{a}+r\right)^2\right]^\frac{1}{2 } } \ , \omega(x , y , z),\end{aligned}\ ] ] where @xmath55 using the following lyapunov - like function @xmath56^{\frac{1}{2}}},\ ] ] where @xmath57 one obtains that @xmath58 therefore , if condition is satisfied and @xmath59 then for system with the given @xmath29 and @xmath38 theorem [ theorem : th1 ] gives estimation . if condition is valid and @xmath60 , then the conditions of theorem [ theorem : th2 ] are satisfied and any solution bounded on @xmath61 tends to an equilibrium as @xmath62 . for system with @xmath44 we can obtain another upper estimation @xcite . [ theorem : th4 ] let @xmath44 , @xmath63 and the following relations hold @xmath64 then @xmath65 we use the same idea but choose the following matrix @xmath66 and the following lyapunov - like function @xmath67 where @xmath68 further it is shown that under conditions of the theorem it is always possible to choose the parameters @xmath69 , @xmath70 such that @xmath71 where @xmath72 hence , for the chosen @xmath29 and @xmath38 the main inequality of theorem [ theorem : th1 ] @xmath73 is satisfied and this completes the proof . [ conseq ] if 1 . @xmath74 , @xmath75 + or 2 . @xmath44 , @xmath63 and @xmath76 then the lyapunov dimension of the zero equilibrium of system with @xmath77 or @xmath78 coincides with . thus for @xmath79 we have @xmath80 note that this exact formula coincides with the formula for the classical lorenz system @xcite . in this section the above results , concerning generalization of lorenz system , are applied to system , describing the convection of fluid within an ellipsoidal rotating cavity . nonsingular linear transformation , obviously , does not change the lyapunov dimension . in general case , it is known ( see , e.g. @xcite ) the invariance of the lyapunov dimension under a diffeomorphism . it is known @xcite that for system if @xmath81 , then there exists one equilibrium @xmath82 . it is stable if @xmath83 or @xmath84 . if @xmath85 , then there exist two additional equilibria @xmath86 and they are stable if @xmath87 or @xmath88 , @xmath89 . here @xmath90 in @xcite by means of numerical simulations for the case when parameter @xmath91 it was found certain values of the parameters @xmath3 and @xmath2 for which in system it is observed either a periodic regime ( i.e. there exists a limit cycle ) or a chaotic regime ( and for system there exists a chaotic attractor ) . in @xcite for parameters @xmath92 in the phase space of system there was obtained a chaotic self - excited attractor and for parameters @xmath93 there was localized numerically a chaotic hidden attractor ( fig . [ fig : attractors ] ) . using linear transformation and relation , one can obtain corresponding parameters for system such that there exists self - excited or hidden attractor . by numerical methods @xcite one can get the following values of lyapunov dimension for the local self - excited attractor of system : @xmath94 and for the local hidden attractor of system : @xmath95 these values correspond to the values of lyapunov dimension for the global attractor @xmath96 of system obtained in corollary [ conseq ] . for ( self - excited local attractor ) we have @xmath97 and for ( hidden local attractor ) we have @xmath98 this work was supported by the russian science foundation ( 14 - 21 - 00041 ) leonov , g. , kuznetsov , n. , and mokaev , t. ( 2015b ) . homoclinic orbit and hidden attractor in the lorenz - like system describing the fluid convection motion in the rotating cavity . , 28(doi:10.1016/j.cnsns.2015.04.007):166174 . leonov , g. a. and kuznetsov , n. v. ( 2013 ) . hidden attractors in dynamical systems . from hidden oscillations in hilbert - kolmogorov , aizerman , and kalman problems to hidden chaotic attractors in chua circuits . , 23(1 ) . 1330002 . semenov , v. , korneev , i. , arinushkin , p. , strelkova , g. , vadivasova , t. , and anishchenko , v. ( 2015 ) . numerical and experimental studies of attractors in memristor - based chua s oscillator with a line of equilibria . noise - induced effects . , 224(8):15531561 .

in the past two decades lyapunov functions are used for the estimation of attractor dimensions . by means of these functions the upper estimate of lyapunov dimension for rssler attractor and the exact formulas of lyapunov dimension for hnon , chirikov , and lorenz attractors are obtained . in this report the simplest model , suggested by glukhovsky and dolzhansky , which describes a convection process in rotating fluid , is considered . a system of differential equations for this model is a generalization of lorenz system . for the lyapunov dimension of attractor of the model , the upper estimate is obtained . chaos , chaotic attractor , generalized lorenz system , lyapunov dimension , lyapunov function , fluid convection

in the classical theory of partitions , compositions were first defined by macmahon @xcite as ordered partitions . for example , there are 5 partitions and 8 compositions of 4 . the partitions are @xmath1 and the compositions are @xmath2@xmath3 . agarwal and andrews @xcite defined an @xmath0-colour partition as a partition in which a part of size @xmath0 can come in @xmath0 different colours . we use a subscript to index the colour and write @xmath4 for the parts of size @xmath0 . analogous to macmahon s ordinary compositions agarwal @xcite defined an @xmath0-colour composition as an @xmath0-colour ordered partition . thus , for example , there are 8 @xmath0-colour compositions of 3 , viz . , @xmath5 & 2_{1}1_{1},~2_{2}1_{1},~1_{1}2_{1},~1_{1}2_{2},\\[.2pc ] & 1_{1}1_{1}1_{1}.\end{aligned}\ ] ] more properties of @xmath0-colour compositions were found in @xcite . @xcite@xmath6 a composition is said to be self - inverse when the parts of the composition read from left to right are identical with when read from right to left . analogous to the above definition of classical self - inverse compositions we define an @xmath0-colour self - inverse composition as follows : @xmath6 an @xmath0-colour composition whose parts read from left to right are identical with when read from right to left is called an @xmath0-colour self - inverse composition . thus , for example , there are 11 @xmath0-colour self - inverse compositions of 5 viz . , @xmath7 & 1_{1}3_{1}1_{1 } , 1_{1}3_{2}1_{1 } , 1_{1}3_{3}1_{1},\\[.2pc ] & 2_{1}1_{1}2_{1 } , 2_{2}1_{1}2_{2},\\[.2pc ] & 1_{1}1_{1}1_{1}1_{1}1_{1}.\end{aligned}\ ] ] @xmath6 the lucas sequence @xmath8 is defined as @xmath9 and @xmath10 when @xmath11 . @xmath6 a point whose co - ordinates are integers is called a lattice point ( unless otherwise stated , we take these integers to be non - negative ) . @xmath6 by a lattice path ( or simply a path ) , we mean a minimal path via lattice points taking unit horizontal and unit vertical steps . @xmath6 the fibonacci sequence @xmath12 is defined as @xmath13 and @xmath14 when @xmath11 . agarwal @xcite proved the following theorem . let @xmath15 and @xmath16 denote the enumerative generating functions for @xmath17 and @xmath18 , respectively , where @xmath17 is the number of @xmath0-colour compositions of @xmath19 into @xmath20 parts and @xmath18 is the number of @xmath0-colour compositions of @xmath19 . then @xmath21 & c(q ) = \frac{q}{1 - 3q+q^{2}},\\[.3pc ] & c(m,\nu ) = \left(\begin{array}{c } \nu+m-1\\[.2pc ] 2m-1 \end{array}\right),\\[.2pc ] & c(\nu ) = f_{2\nu}.\end{aligned}\ ] ] in this section we shall prove the following explicit formulas for @xmath0-colour self - inverse compositions . let @xmath22 be the number of @xmath0-colour self - inverse compositions of @xmath19 into @xmath20 parts . then @xmath23 2m-3 \end{array}\right)\quad { \rm for } \ \ m>1,\\[.3pc ] \hskip -4pc ( { \rm ii})\hskip 2.6pc & a(2m,2\nu ) = \left(\begin{array}{c } \nu+m-1\\[.2pc ] 2m-1 \end{array}\right),\\[.3pc ] \hskip -4pc ( { \rm iii})\hskip 2.6pc & a(2m-1,2\nu ) = \sum_{l=1}^{\nu-1 } 2l\left(\begin{array}{c } \nu+m - l-2\\[.2pc ] 2m-3 \end{array}\right)\quad { \rm for } \ \ m>1.\end{aligned}\ ] ] we shall prove ( i ) and the other parts can be proved similarly . ( i ) obviously , an odd number can have self - inverse @xmath0-colour compositions only when the number of parts is odd . a self - inverse @xmath0-colour composition of an odd number @xmath24 into @xmath25 parts can be read as a central part , say , @xmath26 , and two identical @xmath0-colour compositions of @xmath27 into @xmath28 parts on each side of the central part . the number of @xmath0-colour compositions of @xmath27 into @xmath28 parts is @xmath29 2m-3 \end{smallmatrix}\right)$ ] by ( 1.3 ) . now , the central part can appear in @xmath26 ways and takes values from 1 to @xmath30 . so , @xmath31 varies from 1 to @xmath32 . therefore , for @xmath33 , the number of @xmath0-colour self - inverse compositions of @xmath24 into @xmath34 parts is @xmath35 2m-3 \end{array}\right).\ ] ] for @xmath36 , there will be @xmath24 @xmath0-colour self - inverse compositions . therefore , the total number of @xmath0-colour self - inverse compositions of @xmath37 is @xmath38 2m-3\end{array}\right).\ ] ] in this section , we shall prove the following : if @xmath39 and @xmath40 denote the number of @xmath0-colour self - inverse compositions of @xmath41 , @xmath42 into an even number of parts , @xmath42 into an odd number of parts and @xmath42 ( that is , @xmath43 ) respectively , then the following theorem holds . the sequences @xmath44 satisfy the following recurrences . 1 . @xmath45 and @xmath46 for @xmath47 2 . @xmath48 and @xmath49 for @xmath50 3 . @xmath51 and @xmath52 for @xmath50 4 . @xmath53 and @xmath54 for @xmath55 where @xmath56 theorem 2.1 implies that @xmath57 2m-1 \end{array}\!\right)\quad\!\hbox{and}\quad\ ! c_{\nu } = 2\nu+ \sum_{m=2}^{\nu } \sum_{l=1}^{\nu-1 } 2l \left(\begin{array}{c } \nu+m - l-2\\[.2pc ] 2m-3 \end{array}\right).\end{aligned}\ ] ] \(i ) we have @xmath58 2m-3 \end{array}\right)\\[.3pc ] \hskip -4pc & = ( 2n+1)+ \sum_{m=2}^{n } \sum_{l=1}^{n-1}(2l-1)\left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-3 \end{array}\right)\\[.3pc ] \hskip -4pc & \quad\ , + \sum_{m=2}^{n}(2l-1)\left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-3 \end{array}\right ) \bigg|_{l = n}\\[.3pc ] \hskip -4pc & \quad\ , + \sum_{l=1}^{n } ( 2l-1 ) \left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-3 \end{array}\right)\bigg|_{m = n+1}\\[.3pc ] \hskip -4pc & = ( 2n+1 ) + \sum_{m=2}^{n } \sum_{l=1}^{n-1 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-3 \end{array}\right ) + ( 2n-1)+1\\[.3pc ] \hskip -4pc & = 4n+1+\sum_{m=2}^{n } \sum_{l=1}^{n-1 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-3 \end{array}\right)\\[.3pc ] \hskip -4pc & \quad\ , + \sum_{m=2}^{n } \sum_{l=1}^{n-1 } ( 2l-1)\left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-4 \end{array}\right)\\[.3pc ] \hskip -4pc & = 4n+1+\{a_{n-1}-(2n-1)\ } + \sum_{m=1}^{n-1 } \sum_{l=1}^{n-1 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-2 \end{array}\right)\\[.3pc ] \hskip -4pc & = 2n+2+a_{n-1}+ \sum_{m=2}^{n-1 } \sum_{l=1}^{n-1 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-2 \end{array}\right ) + \sum_{l=1}^{n-1}(2l-1)\\[.3pc ] \hskip -4pc & = 2n+2+a_{n-1}+ \sum_{m=2}^{n-1 } \sum_{l=1}^{n-1}(2l-1 ) \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-3 \end{array}\right)\\[.3pc ] \hskip -4pc & \quad\ , + \sum_{m=2}^{n-1 } \sum_{l=1}^{n-1 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-2 \end{array}\right ) + \sum_{l=1}^{n-1 } ( 2l-1)\\[.3pc ] \hskip -4pc & \quad\ , \left(\hbox{\rm by the binomial identity}\ \left(\begin{array}{c } n\\[.1pc ] m \end{array } \!\right ) = \left(\begin{array}{c } n-1\\[.1pc ] m-1 \end{array } \!\right ) + \left(\begin{array}{c } n-1\\[.1pc ] m \end{array}\!\right)\right)\\ \hskip -4pc\end{aligned}\ ] ] @xmath59 \hskip -4pc & \quad\ , + \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-2\\ 2m-2 \end{array}\right ) \quad + ( n-1)^{2}\\[.3pc ] \hskip -4pc \phantom{a_{n } } & = 2 + 2a_{n-1}+(n-1)^{2}\\[.3pc ] \hskip -4pc & \quad\ , + \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left\{\left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-2 \end{array}\right ) - \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-3 \end{array}\right)\right\}\\[.3pc ] \hskip -4pc & = 2 + 2a_{n-1}+(n-1)^{2}+\sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-1\\[.1pc ] 2m-2 \end{array}\right)\\[.3pc ] \hskip -4pc & \quad\ , - \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-3\\[.1pc ] 2m-3 \end{array}\right ) - \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-3\\[.1pc ] 2m-4 \end{array}\right)\\[.3pc ] \hskip -4pc & = 2 + 2a_{n-1}+(n-1)^{2}+ \left\ { \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-2 \end{array}\right)\right.\\[.3pc ] \hskip -4pc & \quad\ , + \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left . \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-3 \end{array } \right)\right\ } - \{a_{n-2 } - ( 2n-3)\}\\[.3pc ] \hskip -4pc & \quad\ , - \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-3\\[.1pc ] 2m-4 \end{array}\right)\\[.3pc ] \hskip -4pc & = 2a_{n-1 } - a_{n-2 } + ( n-1)^{2 } + ( 2n-1 ) + \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1)\left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-3 \end{array}\right)\\[.3pc ] \hskip -4pc & \quad\ , + \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1)\left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-2 \end{array}\right ) - \sum_{m=2}^{n-1 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-3\\[.1pc ] 2m-4 \end{array}\right)\\[.3pc ] \hskip -4pc & = 2a_{n-1}-a_{n-2}+(n-1)^{2}+(2n-1)\\[.3pc ] \hskip -4pc & \quad\ , + \left\ { \sum_{m=2}^{n } \sum_{l=1}^{n-1 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-3 \end{array}\right ) - 1 - ( 2n-3 ) \right\}\\[.4pc ] \hskip -4pc & \quad\ , + \left\ { \sum_{m=2}^{n-2 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-2 \end{array}\right ) + \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } 2n - l-3\\[.1pc ] 2n-4 \end{array}\right)\right\}\\[.4pc ] \hskip -4pc & \quad\ , -\left\ { \sum_{m=2}^{n-2 } \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n+m - l-2\\[.1pc ] 2m-2 \end{array } \right ) + \sum_{l=1}^{n-2 } ( 2l-1 ) \left(\begin{array}{c } n - l-1\\[.1pc ] 0 \end{array}\right ) \right\}\\[.35pc ] \hskip -4pc & = 2a_{n-1}-a_{n-2}+(n-1)^{2}+(2n-1)+ \{a_{n-1}-(2n-1)-1-(2n-3)\}\\[.3pc ] \hskip -4pc & \quad\ , + 1-(n-2)^{2}\ ] ] @xmath60 \hskip -4pc & = 3a_{n-1}-a_{n-2}.\end{aligned}\ ] ] the other parts can be proved on the same lines . to prove that @xmath61 we split the compositions enumerated by @xmath62 into three classes : 1 . enumerated by @xmath63 and having @xmath64 on both extremes , 2 . enumerated by @xmath63 and having @xmath65 on both extremes , @xmath66 and @xmath67 and @xmath0-colour self - inverse compositions of @xmath41 of the form @xmath68 3 . enumerated by @xmath63 and having @xmath69 on both extremes , @xmath66 , @xmath70 and those enumerated by @xmath71 . we transform the @xmath0-colour compositions in class ( i ) by deleting @xmath64 on both extremes . this produces an @xmath0-colour composition enumerated by @xmath72 . conversely , given any @xmath0-colour composition enumerated by @xmath72 we add @xmath64 on both extremes to produce the elements of the first class . in this way we establish that there are exactly @xmath72 elements in the first class . next , we transform the @xmath0-colour compositions in class ( ii ) by subtracting 1 from @xmath73 , that is , replacing @xmath65 by @xmath74 and subtracting 2 from @xmath75 of @xmath76 . this transformation also establishes the fact that there are exactly @xmath72 elements in class ( ii ) . finally , we transform the elements in class ( iii ) as follows : subtract @xmath64 from @xmath69 , @xmath77 , that is , replace @xmath69 by @xmath78 . we will get those @xmath0-colour self - inverse compositions of @xmath79 whose extremes are @xmath69 , except self - inverse compositions in one part only . also replace @xmath80 by @xmath81 and @xmath82 by @xmath83 , and to get the remaining @xmath0-colour self - inverse compositions from @xmath71 we add 1 to both extremes , that is , replace @xmath65 by @xmath84 . for @xmath0-colour self - inverse compositions into one part we add 2 , that is , replace @xmath85 by @xmath86 , @xmath87 . we see that the number of @xmath0-colour compositions in class ( iii ) is also equal to @xmath72 . so each class has @xmath72 elements . hence , @xmath88 . in this section , we will prove the following theorem . @xmath89 \hskip -4pc { \rm ( ii ) } \hskip 2.6pc & \sum_{n=1}^{\infty}b_{n}q^{n}=\frac{q}{q^{2}-3q+1},\\[.3pc ] \hskip -4pc { \rm ( iii ) } \hskip 2.6pc & \sum_{n=1}^{\infty}c_{n}q^{n}=\frac{2q}{q^{2}-3q+1},~{\rm and}\\[.3pc ] \hskip -4pc { \rm ( iv ) } \hskip 2.6pc & \sum_{n=1}^{\infty}d_{n}q^{n}=\frac{3q}{q^{2}-3q+1}.\end{aligned}\ ] ] @xmath90 & = 1 + 4q+\sum_{n=2}^{\infty}(3a_{n-1}-a_{n-2})q^{n}\\[.2pc ] & = 1 + 4q+3\sum_{n=1}^{\infty}a_{n}q^{n+1}- \sum_{n=0}^{\infty}a_{n}q^{n+2}\\[.2pc ] & = 1 + 4q+3\sum_{n=0}^{\infty}a_{n}q^{n+1 } -3q-\sum_{n=0}^{\infty}a_{n}q^{n+2}\\[.2pc ] \hskip -4pc ( q^{2}-3q+1)\sum_{n=0}^{\infty}a_{n}q^{n } & = 1+q\\[.2pc ] \hskip -4pc \sum_{n=0}^{\infty}a_{n}q^{n}&=\frac{1+q}{q^{2}-3q+1}.\end{aligned}\ ] ] this completes the proof of ( i ) . the proofs of ( ii ) and ( iii ) are similar and hence are omitted . by adding ( ii ) and ( iii ) we get ( iv ) . for @xmath91 , @xmath92 where @xmath93 we will prove the result by induction on @xmath0 . the result is obviously true for @xmath94 . we assume that the result is true for @xmath0 , that is , @xmath95 n - k \end{array } \right ) a_{k}=1.\ ] ] now @xmath96 n+1-k \end{array}\right ) a_{k}\\[.2pc ] & \quad\ , = \sum_{k=0}^{n+1}(-1)^{n+1+k } \left\ { \left(\begin{array}{c } 2n+2\\[.1pc ] n+1-k\end{array } \right)+ \left(\begin{array}{c } 2n+2\\[.1pc ] n - k\end{array } \right ) \right\ } a_{k}\ ] ] @xmath97 n+1-k\end{array}\right)\right . + 2 \left(\begin{array}{c } 2n+1\\[.1pc ] n - k\end{array}\!\right ) + \left.\left(\begin{array}{c } 2n+1\\[.1pc ] n - k-1 \end{array}\!\right)\right\}a_{k}\\[.2pc ] & \qquad\ \left({\rm by}\ \left(\begin{array}{c } n\\[.1pc ] k\end{array}\right)= \left(\begin{array}{c } n-1\\[.1pc ] k\end{array}\right ) + \left(\begin{array}{c } n-1\\[.1pc ] k-1 \end{array}\right)\right)\\[.2pc ] & \quad\ , = \sum_{k=0}^{n+1}(-1)^{n+1+k } \left(\begin{array}{c } 2n+1\\[.1pc ] n+1-k\end{array}\right ) a_{k}-2 + \sum_{k=1}^{n}(-1)^{n+k } \left(\begin{array}{c } 2n+1\\[.1pc ] n - k\end{array}\right ) a_{k-1}\\[.2pc ] & \quad\ , = \sum_{k=0}^{n+1}(-1)^{n+1+k } \left(\begin{array}{c } 2n+1\\[.1pc ] n+1-k \end{array } \right)a_{k}-2 + 3 - 3(-1)^{n } \left(\begin{array}{c } 2n+1\\[.1pc ] n \end{array}\right)\\[.2pc ] & \qquad\ -\sum_{k=2}^{n+1}(-1)^{n+1+k } \left(\begin{array}{c } 2n+1\\[.1pc ] n+1-k \end{array } \right ) a_{k}\\[.2pc ] & \quad\ , = \sum_{k=0}^{1}(-1)^{n+1+k } \left(\begin{array}{c } 2n+1\\[.1pc ] n+1-k \end{array}\right ) a_{k}+1 - 3(-1)^{n } \left(\begin{array}{c } 2n+1\\[.1pc ] n\end{array}\right)\\[.2pc ] & \quad\ , = 1.\end{aligned}\ ] ] therefore the result is true for @xmath98 . hence the result is true for all @xmath99 . for @xmath100 , @xmath101 2m-3 \end{array}\right)\\[.2pc ] \hskip -4pc & \quad\ , = \sum_{m=0}^{\nu } \frac{2\nu+1}{2\nu+1-m } \left(\begin{array}{c } 2\nu+1-m\\[.1pc ] m \end{array}\right),\\[.2pc ] \hskip -4pc \hbox{\rm ( ii ) } \hskip 2.95pc & \nu + \sum_{m=2}^{\nu } \sum_{l=1}^{\nu-1 } l\left(\begin{array}{c } \nu + m - l-2\\[.1pc ] 2m-3 \end{array } \right ) = \sum_{m=1}^{\nu } \left(\begin{array}{c } \nu + m-1\\[.1pc ] 2m-1 \end{array } \right).\end{aligned}\ ] ] @xmath6 \(i ) let @xmath102 and @xmath103 clearly , @xmath104 and we have already proved @xmath105 for @xmath106 . now , for @xmath106 , we have @xmath107 m \end{array}\right)\\[.2pc ] \hskip -4pc & = \sum_{m=0}^{\nu } \left\{\left(\begin{array}{c } 2\nu + 1 - m\\[.1pc ] m \end{array}\right ) + \left(\begin{array}{c } 2\nu - m\\[.1pc ] m-1 \end{array}\right)\right\}\\[.2pc ] \hskip -4pc & \quad\ , \left(\hbox{\rm by the binomial identity}\ \frac{n}{n - m}\left(\begin{array}{c } n - m\\[.1pc ] m \end{array}\right ) = \left(\begin{array}{c } n - m\\[.1pc ] m \end{array}\right ) + \left(\begin{array}{c } n - m-1\\[.1pc ] m-1 \end{array } \right ) \right)\\[.2pc ] \hskip -4pc & = \sum_{m=0}^{\nu } \left\ { \left(\begin{array}{c } 2\nu - m\\[.1pc ] m \end{array } \right)+ \left(\begin{array}{c } 2\nu - m-1\\[.1pc ] m-1 \end{array}\right)\right\ } + b_{\nu-1}\\[.2pc ] \hskip -4pc & \quad\ , \left(\hbox{\rm by the binomial identity}\ \left(\begin{array}{c } n\\[.1pc ] m\end{array}\right ) \right.\\[.2pc ] \hskip -4pc & \qquad\ , \left . = \left(\begin{array}{c } n-1\\[.1pc ] m \end{array } \right ) + \left(\begin{array}{c } n-1\\[.1pc ] m-1 \end{array}\right ) \hbox{\rm and the definition of}\ b_{\nu}\right)\\[.2pc ] \hskip -4pc & = b_{\nu-1}+b_{\nu-1}+ \sum_{m=0}^{\nu } \left\ { \left(\begin{array}{c } 2\nu-1-m\\[.1pc ] m-1 \end{array}\right ) + \left(\begin{array}{c } 2\nu-2-m\\[.1pc ] m-2 \end{array}\right)\right\}\\[.2pc ] \hskip -4pc & = 2b_{\nu-1 } + \sum_{m=0}^{\nu } \left\{\left(\begin{array}{c } 2\nu - m\\[.1pc ] m-1 \end{array}\right ) - \left(\begin{array}{c } 2\nu - m-1\\[.1pc ] m-2 \end{array}\right ) + \left(\begin{array}{c}2\nu-1-m\\[.1pc ] m-2\end{array}\right)\right.\\[.2pc ] \hskip -4pc & \quad\ , -\left . \left(\begin{array}{c } 2\nu-2-m\\[.1pc ] m-3 \end{array}\right)\right\}\\[.2pc ] \hskip -4pc & = 2b_{\nu-1 } + \sum_{m=0}^{\nu } \left\ { \left(\begin{array}{c } 2\nu - m\\[.1pc ] m-1 \end{array } \right)+ \left(\begin{array}{c } 2\nu-1-m\\[.1pc ] m-2 \end{array}\right ) \right\}\\[.2pc ] \hskip -4pc & \quad\ , - \sum_{m=0}^{\nu } \left\ { \left(\begin{array}{c } 2\nu - m-1\\[.1pc ] m-2 \end{array } \right ) + \left(\begin{array}{c } 2\nu-2-m\\[.1pc ] m-3\end{array}\right)\right\}\\[.2pc ] \hskip -4pc & = 2b_{\nu-1}+\sum_{m=0}^{\nu-1}\left\{\left(\begin{array}{c } 2\nu - m-1\\[.1pc ] m \end{array } \right)+ \left(\begin{array}{c } 2\nu - m-2\\[.1pc ] m-1 \end{array}\right ) \right\}\\[.2pc ] \hskip -4pc & \quad\ , - \sum_{m=0}^{\nu-2 } \left\ { \left(\begin{array}{c } 2\nu - m-3\\[.1pc ] m \end{array}\right ) + \left(\begin{array}{c}2 \nu - m-4\\[.1pc ] m-1 \end{array}\right ) \right\}\\[.2pc ] \hskip -4pc & = 2b_{\nu-1}+b_{\nu-1}-b_{\nu-2}\\[.2pc ] \hskip -4pc & = 3b_{\nu-1}-b_{\nu-2}.\end{aligned}\ ] ] hence @xmath108 , for all @xmath100 as these satisfy the same initial conditions and the same recurrence relation . \(ii ) by comparing generating functions for @xmath109 and @xmath110 in 4 and considering the remark below theorem 3.1 , we get ( ii ) easily . the following theorem provides a combinatorial interpretation of theorem 6.1 . @xmath6 1 . for @xmath100 , let @xmath111 denote the number of self - inverse @xmath0-colour compositions of @xmath112 . let @xmath113 , where @xmath114 is the number of lattice paths from @xmath115 to @xmath116 under the following conditions : .3pc 2 . the paths do not have two vertical steps in succession which ensures that they do not cross @xmath117 . 3 . both the first and last steps can not be vertical . + then @xmath118 for all @xmath19 . the number of @xmath0-colour self - inverse compositions of @xmath119 equals three times the number of @xmath0-colour compositions of @xmath19 . @xmath6 1 . in theorem 3.1 we have proved that @xmath111 satisfies the following recurrence relation @xmath120 agarwal @xcite proved that @xmath121 also satisfy the same initial conditions and recurrence relation . this shows that @xmath122 . he also showed that @xmath123 , and it is known that the right - hand side of ( i ) of theorem 6.1 equals @xmath121 ( see p. 185 of @xcite ) . this leads to the first part of the theorem . 2 . from theorem 4.1 , we have@xmath124 . now ( b ) follows from the fact that @xmath125 also equals the number of @xmath0-colour compositions of @xmath19 in view of ( 1.2 ) and theorem 4(ii ) . the authors are thankful to the referee for his suggestions which led to a better presentation of the paper . one of the authors ( gn ) is supported by csir award no . 9/135(468)/2k3-emr - i and the other author ( aka ) is supported by csir research grant no . 25(0128)/02/emr - ii .

macmahon s definition of self - inverse composition is extended to @xmath0-colour self - inverse composition . this introduces four new sequences which satisfy the same recurrence relation with different initial conditions like the famous fibonacci and lucas sequences . for these new sequences explicit formulas , recurrence relations , generating functions and a summation formula are obtained . two new binomial identities with combinatorial meaning are also given . [ theoree]*theorem * [ theoree]_proof of theorem _

the asian summer monsoon is the largest single abnormality in the global climate system ( shukla , 2007 ) . the seasonal rainfall brought by the southwest indian summer monsoon supplies 80% of southeast asia s annual precipitation and is vital to sustaining the region s agriculture which supports nearly a quarter of the world s population ( sinha et al . , 2007 ) . indian summer monsoon is one of the main weather systems on earth and variations in its intensity have broad economic effects . it has been the most important climate event in india . rainfall over india is subject to a high degree of variations leading to the occurrence of extreme monsoon rainfall deficient ( drought ) or excess ( flood ) over extensive areas of the country . floods and droughts result in many losses of lives , crops etc . ; these play havoc to indian economy and society . cause of abnormal variabilities in monsoon rainfall ( floods and droughts ) is not completely understood . consequently , accurate prediction of rainfall and its variability during monsoon season has been a challenging task . thus , there is greater need to understand the nature and variability of monsoon climatic conditions , especially , whether there is any extra - terrestrial influence ( e.g. cosmic ray variability ) in addition to natural terrestrial climatic conditions . more specifically , it is important to know whether indian monsoon rainfall is significantly influenced by changes in cosmic ray flux , and whether climate cooling is an effect of cosmic ray flux change . if so , then the possible physical mechanism(s ) must be identified . it is well known that cosmic ray flux varies in anti - phase with solar activity over all time scales . on the longer time scales ( millennial , centennial and multi - decadal ) , a number of studies have suggested solar / cosmic ray variability influence on the intensity of monsoonal rainfall in tropical and sub - tropical regions with conflicting results . for example , low rainfall in india coinciding with low solar activity ( or high cosmic ray intensity ) ( e.g. agnihotri et al . , 2002 ; tiwari et al . , 2005 ; gupta et al . , 2005 ; yadava and ramesh , 2007 ) and in north africa and south oman ( neff et al . , 2001 ) . these results imply that increased galactic cosmic ray ( gcr ) intensity is associated with a weakening of the monsoon ( decreased rainfall ) ( kirkby , 2007 ; singh et al . , 2011 ) . in contrast , low rainfall in equatorial east african ( e.g. verschuren et al . , 2000 ) , weaker chinese monsoon ( hong et al . , 2001 ) , and low tropical rainfall in gulf of mexico region have been observed , during high solar activity ( or low cosmic ray intensity ) . occurrence of periods of enhanced monsoonal precipitation in india slightly after the termination of the wolf , sporer and maunder minima periods ( low solar activity / high cosmic ray intensity ) have been reported by khare and nigam ( 2006 ) . this finding is in agreement with the finding of earlier workers , who reported high lake levels from mono lake and chad lake in the vicinity of solar minima ( cosmic ray maxima ) as well as the nile river in africa ( ruzmaikin et al . , 2006 ) . thus there are evidences , although sometimes contrary in nature , that suggest for some cosmic ray influence on monsoon rainfall on multi - decadal , centennial and millennial time scale . on shorter time scales ( decadal to inter annual ) too , solar activity / cosmic ray intensity influence on the rainfall changes in indian summer monsoon have been suggested , but with conflicting results ( e.g. see jagannathan and bhalme , 1973 ; bhalme et al . , 1981 ; hiremath and mandi , 2004 ; bhattacharya and narasimha , 2005 ; badruddin et al . , 2006 , 2009 ) . understanding the factors that control ismr onset , its variability and intensity are highly desired . in particular , it is extremely important to know about the role of extra - terrestrial sources ( e.g. cosmic rays ) in initiating and/or influencing the intensity of rainfall directly ( e.g. by changing the collision / coalescence efficiency in rain clouds ) or indirectly ( e.g. by altering the low cloud amount ) . it is particularly important to search for connection , if any , between the extreme deficiency ( droughts ) or excess ( floods ) in indian summer monsoon rainfall and cosmic ray flux variability during the same ismr periods , even though it is widely accepted that indian monsoon onset and intensity are controlled by large scale atmospheric ( e.g. land - sea temperature contrast ) and global features ( e.g. enso , qbo etc . ) . several studies have shown that the warm phase ( el nino ) is associated with weakening of indian monsoon with overall reduction in rainfall while the cold phase ( la nina ) is associated with the strengthening of the indian monsoon with enhancement in rainfall ( e. g. , sikka , 1980 ; pant and parthasarathy , 1981 ; rasmusson and carpenter , 1983 ) . all the el nino events during 1958 - 1988 were reported to be droughts and all the la nina events were associated with excess ismr . however , weakening of enso - ismr relationship after 1988 were reported in later studies ( e. g. , kripalani and kripalani , 1997 ; kumar et al . , 1999 ; ashok et al . , 2001 ; kripalani et al . , 2003 ) . further , for the 14 consecutive years beginning with 1988 ( 1988 to 2001 ) , there were no droughts , despite the occurrence of el nino ( gadgil et al . , 2004 ) . although 9 out of 12 drought years identified by us can be associated with el nino events , and 9 out of 12 flood years with la nina events , there are reports ( kumar et al . , 2002 ) that out of 22 large negative ismr anomalies that occurred during 1871 - 2001 , only 11 were associated with el nino , while out of 19 large positive ismr anomalies that occurred during the same period , only 8 were associated with la nina . therefore , large deficient / excess ismr does occur in the absence of el nino / la nina and we do not yet understand adequately the response of monsoon to el nino ( gadgil et al . , 2004 ) . thus , there is the possibility of drought / floods in india being influenced by other external agents also . cosmic rays are the only source of ion production in the lower atmosphere . there are suggestions that cosmic ray flux variability may influence the earth s climate also . in view of these suggestions , although controversial , it will be interesting to search for any possibility of a link between gcr and rainfall variability . variations in precipitation potentially caused by changes in the cosmic ray flux have implications for the understating of the cloud and water vapour feedbacks . it is possible that any particular ( e.g. indian ) climate system is more sensitive to smaller variations in cosmic ray intensity than the other . the purpose of this investigation is to determine the relationship , if any , between the indian extreme weather ( drought / flood ) and cosmic ray flux variability . we analysed the gcr flux data to evaluate the possible existence of empirical evidence between cosmic ray variability and precipitation in india during monsoon season . for this study we utilized the gcr fluxes as recorded through the ground based neutron monitors , and perform analysis to look for any possibility of changes in pattern in indian rainfall , in particular , due to variations in gcr flux . for this purpose we adopt the methods of superposed epoch analysis ( singh and badruddin , 2006 ) and regression analysis . we find evidence for a possibility that gcr flux variability may have some influence in suppressing / enhancing the rainfall depending upon the decreasing / increasing nature of gcr variability , in favourable climatic conditions . in this work we adopt an approach that assumes that the rainfall changes can occur only with gcr changes if environmental conditions are suitable , and considering that the rate of gcr flux change , and not the mean gcr flux , may be the key ( laken et al . , 2010 ) . usoskin ( 2011 ) , in a recent review , concluded that it is not the intensity of cosmic rays but its variability that may affect climate . the gcr flux is provided by neutron monitors , which record neutrons generated chiefly by the primary cosmic ray protons that ionize the lower stratosphere and upper troposphere ( venkatesan and badruddin , 1990 ; bazilevskaya and svirzhevskaya , 1998 ) . continuous records of high quality cosmic ray intensity data , measured by neutron monitors located at different latitudes and longitudes on the earth s surface are available from 1964 onwards till date . reliable and good quality data of monsoon rainfall in india are also available for the period 1964 - 2011 ( see table 1 ) and many more years before that , at indian institute of tropical metrology , pune ( india ) website ( http://www.tropmet.res.in/ ) . for this work , we have considered the 48-year period ( 1964 - 2011 ) for which both the gcr intensity and indian summer monsoon rainfall ( ismr ) data are available . a variation of @xmath0 in yearly mean gcr flux over a period of one solar cycle in anti - phase with the @xmath111-year sunspot polar activity cycle is a well - observed known phenomena ( see fig . 1 ) . a number of studies have utilized this change in gcr intensity over solar cycles to suggest ( or refute ) a possible connection between cosmic rays and climate ( clouds , rainfall , temperature etc . ) . however , significant changes / fluctuations in gcr intensity are observed when the data is averaged over monthly and daily time resolutions . at times , with these time resolutions , the gcr flux is observed to increase / decrease by a large amount ( a few percent ) during some months in the same year and during several days in the same month . the purpose of this paper to search the influence , if any , of gcr flux change on the summer monsoon rainfall in india , at regional and seasonal or even shorter time scales . for this purpose out of 48 years from 1964 - 2011 ( see table 1 ) we first identify @xmath1one - forth ( 12 ) years with lowest rainfall in four indian summer monsoon months ( june - september ) ( see table 2 ) . we call them deficient rainfall ( drought ) years and the same number of years ( 12 ) with the highest rainfall in summer monsoon months ( see table 2 ) . we call them heavy rainfall ( flood ) years . in fig . 2 , the deviation of monthly precipitation during superposed ` drought ' and ` flood ' years from superposed average precipitation of 48 years ( 1964 - 2011 ) is shown . in fig . 1 , we have plotted yearly average gcr intensity as observed by neutron monitors located at two different latitudes and longitudes ( see table 3 ) , namely oulu ( http://cosmicrays.oulu.fi/ ) and newark ( http://neutronm.bartol.udel.edu/ ) . unfortunately , there is no neutron monitor located in india whose data for the period 1964 - 2011 can be utilized for this analysis . however , the time variation shown in fig . 1 at two locations on the earth is similar in nature globally with different amplitudes at different latitudes . that is , the nature of gcr variations observed at globally distributed monitors are similar , only differing in amplitudes . however , there are some suggestions ( e.g. eroshenko et al . , 2010 ) that the rainfall and the humidity influence the incoming particle flux around the detector ; moisture around the detector lowers both the neutrons incident to the surface and albedo neutrons . 1 , the drought and flood years are indicated by dashed and dotted vertical lines respectively . from fig . 1 we see that in india droughts / floods can occur at any level of mean gcr flux , minimum / maximum or intermediate level . in other words , these floods / droughts can occur at maximum / minimum / increasing / decreasing phases of the solar activity cycle . thus if we assume mean gcr flux to be the key , then we can conclude that there is no influence of gcr flux on the indian monsoon rainfall of inter - annual scale . next , we proceed to search for any possible influence of gcr flux variability on indian monsoon rainfall during the same period , assuming that it is more likely that the rainfall changes occur only with gcr flux changes if environmental conditions are suitable , and that not the mean gcr flux , but its variability may affect the rainfall amount / climate ( see laken et al . , 2010 ; usoskin , 2011 ) . the cosmic ray count rate has solar cycle dependence , so we normalized the count rate before performing the superposed epoch analysis . each year s data is normalized to the yearly average for that year , then the data is converted to percentage , which in turn allows for a direct comparison of the different data ( i.e. gcr intensity , sunspot number , 10.7 cm solar radio flux , total solar irradiance ) . frist , we have calculated the yearly average for individual years , then each months data is converted into percentage by takeing yearly average as reference . monthly resolution cosmic ray count rate of oulu nm during drought and flood years , mean count rate of individual months with standard deviation ( @xmath2 ) and standard error of mean ( sem = @xmath3 ) of both before and after normalization are tabulated in tables 4a and 4b . we then perform the superposed epoch analysis to study the rate of gcr flux variability during ismr months ( june - september ) averaged over 12 drought years and 12 flood years separately . for this purpose we have utilized the normalized gcr intensity data of three neutron monitors located at different positions on the earth , namely oulu ( finland ) , kiel ( germany ) and newark ( usa ) ( see table 3 ) . these three location data have been analysed to show that the nature of variation is globally similar , only differing in amplitudes . in fig . 3(a ) we have plotted the superposed epoch results of monthly averaged normalized gcr intensity data for deficient rainfall years as observed by oulu neutron monitor count rate . we see that gcr intensity is decreasing during ismr ( june - september ) period ( shaded ) . the rate of decrease has been calculated by fitting a linear curve ( see table 5 ) , taking the pre - monsoon ( may ) value as a reference . the best - fit result shows that the gcr count rate decreases ( negative slope ) with linear correlation coefficient r = @xmath4 ( see table 5 and fig . as the gcr intensity may fluctuate to a large extent on a day - to - day basis , we have done the superposed epoch analysis of the daily normalized gcr count rate , as observed by the oulu neutron monitor , for the same 12 deficient rainfall years . the result of this analysis is plotted in fig . 4(a ) and tabulated in table 5 . we see a continuously decreasing gcr intensity during the summer monsoon period . we did a linear regression to this averaged data considering the pre monsoon ( may ) data as the reference . the best - fit line with negative slope ( r = @xmath5 ) is also shown ( see table 5 and fig . 4 ) . in both monthly and daily cases , we note that the regression line is steeper than it would be if the regression line is obtained using the entire year . to show that such a variation is not confined to one location but its nature is global , we did a similar superposed epoch analysis and best fit linear regression , as earlier , for two more neutron monitor stations data namely kiel and newark using monthly average gcr count rate as well as daily count rate ( see table 5 ) . we see a similar decreasing trend at these locations also . thus , we can infer that the trends of rate of change in gcr flux will be similar in nature at indian locations also . next , we consider the same number ( 12 ) of heavy rainfall years and did a similar superposed epoch analysis of gcr count rate ( both monthly and daily ) data for the same three neutron monitors . we also did a linear regression of these data for the ismr period taking pre - monsoon ( may ) value as a reference . we find that gcr flux is increasing during ismr periods . the best - fitted linear curves with positive slope and correlation coefficients are clearly evident on all three - neutron monitor stations and at both time resolutions ( see figs . 3 , 4 and table 5 ) . the linear regression shows a line with positive slope [ see figs . 3(e ) and 4(e ) ] , and from these figures it is clear that the slope will be less rapid if a regression line is drawn for the entire year . in addition to cosmic rays , we extended our analysis to solar activity parameters , such as sunspot number ( ssn ) and 10.7 cm solar radio flux ( http://omniweb.gsfc.nasa.gov/ ) . the ssn is the oldest directly observed solar activity on the photosphere and a very useful indicator of solar activity . the 10.7 cm solar radio flux is an indicator of activity in the upper chromosphere and lower corona . we also considered total solar irradiance ( tsi ) data ; however , this data is available only from 1979 onwards ( http://www.ngdc.noaa.gov/ ; http://lasp.colorado.edu/home/ sorce / data / tsi - data/ ) . we analysed both monthly and daily resolution data of solar parameters ssn , 10.7 cm solar radio flux and tsi for drought and flood years . we carried out superposed epoch analysis after normalizing the data as done for gcr count rate . superposed epoch results of monthly averaged normalized ssn , 10.7 cm solar radio flux and tsi are plotted in fig . 3 for both drought years [ fig . 3(b - d ) ] and flood years [ fig . 3(f - h ) ] . the rate of change has been calculated by fitting a linear curve , taking the pre - monsoon ( may ) value as a reference . the best - fit results with linear correlation coefficient are also shown ( see fig . 3 and table 5 ) . we have also done the superposed epoch analysis of the daily normalized solar parameters data , for the same 12 drought ( ( b - d ) of fig . 4 ) and flood ( ( f - h ) of fig . 4 ) rainfall years . the rate of change has been calculated by fitting a linear curve to the data . the best - fit results with linear correlation coefficient are tabulated in table 5 . we can see the difference in nature of variability ( slopes ) in gcr flux and solar parameters ( ssn , 10.7 cm solar radio flux , tsi ) , errors in slopes and correlation coefficients ( r ) during drought and flood periods , favouring gcr flux - rainfall relationship . noticeable difference seen in gcr variability is not so clear in solar parameters considered here . however , it is possible that restricting the correlation analysis to ismr months only reduces the apparent dependence on solar indices such as f10.7 and ssn . but the anti - correlation between gcr flux and ssn and f10.7 becomes clearer if the analysis interval is extended to whole year . we observe that on an average , the gcr flux is decreasing during ismr months ( june - september ) with deficient monsoon rainfall ( drought ) in india . on the other hand , gcr flux is increasing during ismr period with heavy rainfall ( flood ) in india . as regards the change in temperature with rainfall changes , we find that there is a strong inverse relation between the rainfall and temperature ( see fig . 5 ) , at least during ismr period . in the view of the results shown in figs . 3 and 4 , there is urgent need to quantify the extent of influence , and to identify the physical mechanism(s ) , responsible for influencing the indian monsoon rainfall through the cosmic ray flux variability . although , we found definite trends i.e. , on the average , heavy rainfall ( floods ) in india occurs during ismr period when gcr flux is increasing in the same season , and gcr flux is decreasing in ismr months during deficient rainfall ( drought ) in indian summer monsoon period . however , a caveat must be added here ; that the rainfall changes can occur with gcr changes only if environmental conditions ( to be identified ) are suitable . this caveat implies that similar trends in rainfall changes with gcr flux changes ( i.e. , deficient rainfall associated with decreasing gcr flux and heavy rainfall associated with increasing gcr flux ) may not be observed at all geographic locations from equator to pole and in all seasons simultaneously , although the nature of gcr flux change is overall similar at almost all locations on the earth only differing in magnitude . moreover , there may be exceptions in india even during ismr season due to unsuitable environmental conditions . as regard the breakdown of the enso - ismr connection after 1988 , mentioned earlier , the breakdown may be because the ismr is less variable . similary to enso , the gcr flux variability has similar properties on both the quiet interval from 1989 to 2002 and the drought and flood periods before and after this gap . thus the breakdown in the ismr - enso connection is not necesserly the evidence for the ismr - gcr hypothesis . the possibility of both the enso and gcr variability contributing to ismr variability in their own way can not be ruled out at this stage . more efforts and rigorous analyses are required to discriminate between enso and gcr as a cause of ismr variability . and maximum temperature [ ( tmax)jjas ] during ismr months . ] most of the studies that attempt to study possible gcr - cloud - climate relationship are focused on longer time scales ( millennial , centennial , multidecadal and decadal ) ( e.g. see reviews by carslaw et al . , 2002 ; kirkby , 2007 ; singh et al . , 2011 ; rao , 2011 ; and references therein ) . however , on shorter time scales too ( inter - annual , seasonal and even smaller ) attempts have been made to search for this relationship with conflicting results . forbush decreases are sudden decreases @xmath1a few percent in cosmic ray intensity within about a day and recover to its pre - reverse level within a week or so ( e.g. see rao , 1972 ; venkatesan and badruddin , 1990 ; kudela , 2009 for reviews on cosmic ray variations at different time scales ) . these forbush decreases in cosmic rays are thought to be an important laboratory for testing possible cosmic ray climate connection . decreases in rainfall in the former soviet union have been reported in the days of the forbush decreases ( stozhkov et al . , 1995 ) . however , most of the eastern mediterranean stations present higher probabilities for a precipitation episode one day after a forbush decrease ( mavrakis and lykoudis , 2006 ) . precipitation changes in relation to gcr flux changes in a short time scale have also been studied by kniveton and todd ( 2001 ) . a recent claim that forbush decreases affect atmospheric aerosol and cloud ( svensmark et al . , 2009 ) has been challenged by other studies ( kulmala et al , 2009 ; laken et al . , 2009 ; calogovic et al . , 2010 ) who found no connection between cosmic rays , aerosols and clouds . however , a more recent study ( dragic et al . , 2011 ) from an analysis of european region data supports the idea that cosmic rays influence the atmospheric process and climate . earlier too , the claim of decreased cloudiness detected during forbush decreases ( pudovkin and veretenenko , 1995 ) was not observed by palle and butler ( 2001 ) even during the same forbush decreases . moreover , the underlying physics suggesting for a connection between cosmic rays , aerosols and cloud is still highly speculative ( legras et al . , 2010 ) and empirical evidences for cosmic ray - cloud relation is still inconclusive ( usoskin , 2011 ) . it was concluded in a review ( kirkby , 2007 ) , based on the available results for longer timescales ( millennial , centennial and multi decadal ) , that increased gcr flux appears to be associated with a cooler climate and a weakening of the monsoon ; and a decreased gcr flux is associated with a warmer climate and strengthening of the monsoon . from our observations on a much shorter time scale during monsoon season in india , we observe that a decreasing gcr flux corresponds to decreasing rainfall and increasing gcr flux corresponds to increasing rainfall . moreover , our preliminary results reported earlier ( badruddin et al . , 2006 , 2009 ) show that temperature and rainfall changes show an opposite behaviour , i.e. , temperature is enhanced during deficient ismr periods and it is lower in heavy rainfall ismr periods . the observation that the cosmic ray intensity is decreasing during ismr months in almost all the years which , are deficient in rainfall ( drought years ) may be interpreted to suggest that a gcr -rainfall relation is possible in indian sub - continent during ismr periods , at least . thus , the gcr - rainfall relation should be considered as a potentially important driver of climate variability . a significant part of precipitation that falls in the tropics is warm rain formed by coalescence of cloud droplets ( kostinski and shaw , 2005 ) . formation of cloud droplets requires a water vapour super saturation environment and particles able to act as cloud condensation nuclei . usually cloud droplets are formed on aerosol particles containing a certain stable fraction . after condensation droplets grow by vapour diffusion and droplet - droplet collision ( coalescence ) , the latter providing more rapid growth as droplet size increases ( harrison and ambaum , 2009 ) . electrical effects play an important role in cloud microphysics . both condensation and coalescence can be influenced by the charge ( pruppacher and klett , 1997 ; tinsley , 2008 ) . early laboratory studies found that raindrops ( of around 0.5 mm diameter ) are about factor 100 more efficient at collecting aerosols when they are charged rather that neutral ( barlow and latham , 1983 ) . grover and beard ( 1975 ) calculated collision efficiencies and found a significant increase in collision efficiency when the droplets were loaded with a charge of the magnitude typical of thunderstorm clouds . khain et al . ( 2004 ) from their simulation results have shown that the injection of just a small fraction of charged particles rapidly triggered the collision process and lead to raindrop formation a few minutes after the injection , thus seeding with charged particles may be a very efficient tool for rain enhancement , they suggested . the collision efficiencies highly depend on droplet charge and size . the collision efficiency is much enhanced in the case of a charged droplet collisions than in case of neutral droplet collisions . more specifically , they found that the collision efficiency between charged and neutral droplets , as well as between droplet charges of opposite polarity , is many orders higher than in the case of gravity - induced collisions . thus , efficient collision takes place between cloud droplets and coalescence to large droplet is enhanced by electrical forces between charged droplets . this significantly increases the rate of raindrop formation ( khain et al . , 2004 ) . another potential mechanism may operate through electrostatic image forces . because of electrostatic image forces , electrical forces between charged droplets are always attractive at small separations whatever the relative polarities of the colliding particles ( tinsley , 2008 ) . in this case , the attraction between droplets may lead to droplet size increase . as the droplet size increases , the droplet - droplet collision ( coalescence ) will lead to more rapid growth , leading to enhanced rate of raindrop formation , as suggested by harrison and ambaum ( 2009 ) . this mechanism appears more likely as the rate of change of gcr flux and not the amount of gcr flux is considered to be the key factor . however , more simulations and experiments need to be performed to demonstrate clearly how the increasing gcr flux corresponds to increasing rainfall and vice versa . thus it is expected that in proper atmospheric / environmental conditions ( e.g. air humidity , aerosols , temperature , cloud type etc . ) increasing gcr flux will increase coalescence efficiency that will lead to bigger rain droplets while decreasing gcr flux will decrease the coalescence efficiency and will suppress the droplet growth . during decreasing flux of cosmic rays , levitation / dispersion of low clouds due to electrical effects ( levin and ziv , 1974 ) may also play some role in such a way that it disperses the low cloud amount in proper climatic conditions . we suspect that in suitable environmental conditions , charge particle ( cosmic ray ) flux rate change modulates the droplet collision and coalescence efficiency and affects the rainfall to certain extent . rain formation is a function of different parameters of macro- and micro - physics . the important parameter for the microphysics is the ambient temperature where clouds reside and formation of raindrops occurs due to process of spontaneous coalescence and accretion ( rogers and yau , 1989 ) . although the initiation of raindrop coalescence remain an unsolved problem in cloud physics ( kostinski and shaw , 2005 ) , we suspect that charge induced cloud microphysics , for example , accelerating / decelerating coalescence to larger raindrops ( harrison and ambaum , 2009 ) is the likely effect that plays some role in affecting the rainfall variability in india during indian summer monsoon season , at least , depending on the increasing / decreasing rate of change of charge particle ( cosmic ray ) flux in the corresponding period , under suitable environmental conditions ( e.g. cloud type , temperature , pressure etc . ) . monsoon rainfall variability is connected with global precipitation ( hulme et al . , 1998 ) . there is a strong inverse relationship between the monsoon variability and tropical belt temperature ( see dugam and kakade , 1999 ; badruddin et al . , 2006 , 2009 ) . considering that change in monsoon rainfall variability is also consistent with the change in global mean precipitation ( hulme et al . , 1998 ) and precipitation / rainfall is inversely related to temperature ( see fig . 8) , we suspect that the monsoon rainfall variability may have some influence on the changes in global temperature also . thus , it should be clarified whether monsoon / rainfall variability plays any role in global warming or its effects are only local . it has been suggested ( ban - weiss et al . , 2011 ) that evaporated water helps in cooling earth as a whole and not just the local area of evaporation . on the other hand , reduction in evaporated water is likely to contribute to global warming significantly . the possible influence of gcr on clouds is a controversial issue . it appears that gcr flux variability plays an important role in influencing the ismr in this season , at least . it is likely , as we suspect , that the physical state of the cloud droplets may play a significant role . local physical ( cloud type , temperature , humidity etc . ) and chemical conditions may play a major role ( enghoff et al . , 2011 ; duplissy et al . , 2010 ; kirkby et al . , 2011 ) in deciding the extent of the influence . physics of liquid and ice cloud may differ ( geirens and ponater , 1999 ) . low clouds generally consist of liquid water droplets ( marsh and svensmark , 2000 ) . it has been suggested that low cloud liquid droplets over the tropics are more sensitive to cosmic ray variability ( palle and butler , 2000 ) . it is suggested that such studies ( i.e. , effects of gcr flux variability on rainfall variability ) on regional hydrological regions have to be studied in more detail . it is also suggested that proper environmental condition in which the influence of gcr flux variability of rainfall is more significant , needs to be identified . although amount of cloud may be dependent on gcr flux , in our hypothesis , we do not consider a direct relationship between the gcr flux and amount of cloud cover as the key ; it is still controversial . we propose an alternate scenario , although speculative , in which the cosmic ray variability influences the rainfall from clouds that are formed in proper environmental conditions . we propose that increasing / decreasing gcr flux influences the rainfall which later results in enhanced / reduced evaporation . this change in evaporation from the earth surface influences the low cloud amount which in turn alters the planetary albedo and concequently there is change in the temperature . however , such influence is only regional or has global effect needs to verified . therefore , more research is needed to understand the relationship amoung variability in gcr , ismr , surface evaporation , low cloud , planetary albedo and temperature . also model studies are needed to understand the extent to which such variability influences the regional and global rainfall and temperature . we find that the decreasing cosmic ray flux does play a role in such a way that the rainfall over this region of the globe ( india ) , at least , is reduced when cosmic ray flux is decreasing . we speculate that the hypothesis , proposed here , on the basis of indian climate data , can be extended to whole tropical and sub - tropical belt , and that it may contribute to global temperature in some way . in conclusion , a gcr - ismr link seems plausible and the gcr - rainfall relation should be considered as a potentially important driver of climate variability . however , further studies are required to improve our understanding of the link between cosmic rays and summer monsoon climate over india . it is also required to fully investigate the contributions of possible mechanisms , discussed here , to the variability in precipitation . further , once our hypothesis is confirmed , there is an urgent need to identify the local physical and chemical conditions conducive for significant effect of gcr flux variability in influencing the rainfall / precipitation . we suggest the following scenario , although speculative , for possible relationship between gcr flux - rainfall - temperature . 1 . increasing gcr flux @xmath6 increasing rainfall @xmath6 enhanced surface evaporation @xmath6 increased low cloud @xmath6 more scattering of solar radiation back to space ( more planetary albedo ) @xmath6 lower temperature . 2 . decreasing gcr flux @xmath6 decreasing rainfall @xmath6 decreased surface evaporation @xmath6 reduced low cloud @xmath6 less scattering of solar radiation back to space ( less planetary albedo ) @xmath6 higher temperature . we thank station manager ilya usoskin and sodankyla geophysical observatory for the online availability of oulu neutron monitor data , the national science foundation ( supporting bartol research institute neutron monitors ) and principle investigator john w. bieber for the online availability of newark neutron monitor data and christian t. steigies and extraterrestrial physics department of university of kiel for the online availability of kiel neutron monitor data . availability of indian climate data through indian institute of tropical meteorology pune s website and its use is gratefully acknowledged with thanks . we also acknowledge the use of ssn and 10.7 cm solar radio flux data available through the nasa / gsfc omni web interface , total solar irradiance data through national geophysical data center website and sorce homepage . the authors also thank the editor and referees , whose comments and suggestions helped us to improve the paper . badruddin , singh , y. p. , singh , m. , 2006 . does solar variability affect indian ( tropical ) weather and climate ? : an assessment . in : gopalswamy , n. , bhattacharya , a. ( eds . ) , solar influence on the heliosphere and earth s environment : recent progress and prospects ( proc . ilws workshop ) . quest publications , 444 - 447 . gierens , k. , ponater , m. , 1999 . comment on variation of cosmic ray flux and global cloud coverage - a missing link in solar - climate relationships by h. svensmark and e. friis - christensen ( 1997 ) . 61 , 795- 797 . grover , s. n. , beard , k. v. , 1975 . a numerical determination of the efficiency with which electrically charged cloud drops and small raindrops collide with electrically charged spherical particles of various densities . 32 , 2156 - 2165 . hong , y.t . , wang , z. g. , jiang , h. b. , lin , q. h. , hong , b. , zhu , y. x. , wang , y. , xu , l. s. , leng , x. t. , li , h. d. , 2001 . a 6000-year record of changes in drought and precipitation in northeastern china based on a 13c time series from peat cellulose . earth planet . 185 , 111 - 119 . khain , a. , pokrovsky , a. , pinsky , m. , seifert , a. , phillips , v. , 2004 . simulation of effects of atmospheric aerosols on deep turbulent convective clouds using a spectral microphysics mixed - phase cumulus cloud model . part i : model description and possible applications . 61 , 2963 - 2982 . kulmala , m. , asmi , a. , lappalainen , h. k. , carslaw , k. h. , et al . , 2009 . introduction : european integrated project on aerosol cloud climate and air quality interactions ( eucaari ) - integrating aerosol research from nano to global scales . chem . phys . 9 , 2825 - 2841 . pruppacher , h. r. , klett , j. d. , 1997 . microphysics of clouds and precipitation , ( second revised and enlarged edition with an introduction to cloud chemistry and cloud electricity ) , kluwer academic publishers , dordrecht , pp 954 . sikka , d. r. , 1980 . some aspects of the large - scale fluctuations of summer monsoon rainfall over india in relation to fluctuations in the planetary regional scale circulation parameters . indian acad . sci . earth planet . 89 , 179 - 195 . sinha , a. , cannariato , k. g. , stott , l. d. , cheng , h. , edwards , r. l. , yadava , m. g. , ramesh , r. , singh , i. b. , 2007 . a 900-year ( 600 to 1500 a.d . ) record of the indian summer monsoon precipitation from the core monsoon zone of india . 34 , l16707 .

we study the role of galactic cosmic ray ( gcr ) variability in influencing the rainfall variability in indian summer monsoon rainfall ( ismr ) season . we find that on an average during drought ( low ismr ) periods in india , gcr flux is decreasing , and during flood ( high ismr ) periods , gcr flux is increasing . the results of our analysis suggest for a possibility that the decreasing gcr flux during the summer monsoon season in india may suppress the rainfall . on the other hand , increasing gcr flux may enhance the rainfall . we suspect that in addition to real environmental conditions , significant levitation / dispersion of low clouds and hence reduced possibility of collision / coalescence to form raindrops suppresses the rainfall during decreasing gcr flux in monsoon season . on the other hand , enhanced collision / coalescence efficiency during increasing gcr flux due to electrical effects may contribute to enhancing the rainfall . based on the observations , we put forward the idea that , under suitable environmental conditions , changing gcr flux may influence precipitation by suppressing / enhancing it , depending upon the decreasing / increasing nature of gcr flux variability during monsoon season in india , at least . we further note that the rainfall variability is inversely related to the temperature variation during ismr season . we suggest an explanation , although speculative , how a decreasing / increasing gcr flux can influence the rainfall and the temperature . we speculate that the proposed hypothesis , based on the indian climate data can be extended to whole tropical and sub - tropical belt , and that it may contribute to global temperature in a significant way . if correct , our hypothesis has important implication for the sun - climate link . sun - earth connection , galactic cosmic rays , summer monsoon rainfall , temperature

the equations of point vortices motion in the plane were introduced by helmholtz @xcite and described as an hamiltonian system by kirchhoff in 1876 @xcite . these equations were first generalized to describe the motion of point - vortices on a sphere by bogolmonov in 1977 @xcite . bogomolnov s work was put into solid mathematical grounds by kimura and okamoto in 1987 @xcite . independently hally wrote in 1979 @xcite the equations for vortices motion in symmetric surfaces of revolution ( which englobe the sphere ) . to write his equations hally represented the surface @xmath2 on the complex plane with a metric conformal to the euclidean metric : let the @xmath3-th vortex , with vorticity @xmath4 be represented by the complex coordinate @xmath5 . calling the surface s conformal factor @xmath6 , hally s equations are @xmath7 the difficulty in working with these equations is that one must know explicitly the conformal factor @xmath6 and this factor is not known for many surfaces but the sphere . the aim of this paper is to write the equations for the n - vortex problem on the ellipsoid of revolution @xmath8 as an @xmath9-series when @xmath10 . to achieve this goal we write a perturbative coordinate transformation from the ellipsoid into the plane ( see ( [ trans]))and impose conformality at each order . it turns out that the transformation s n - th order term is given by the solution of a certain linear differential equation and that the first order term can be computed explicitly ( see ( [ equapert ] ) ) . using hally s equations we calculated rigorously the first order perturbation term for the vortex equations on the ellipsoid . in spherical coordinates the truncated first order equations are given by the hamiltonian system with hamiltonian @xmath11 where @xmath12 \ , , \ ] ] @xmath13\ ] ] and symplectic two - form @xmath14 the paper is organized as follows : in section ( [ hallysec ] ) we introduce all the concepts used in the derivation of the vortex equations . hally s equations are written as an invariant hamiltonian system and , for the spherical case , are shown to be equivalent to bogolmonov s equations . in section ( [ pertubative ] ) we derive the expression for the first order perturbation term . the main idea is to write the conformal transformation from the ellipsoid into the plane as a formal series in the eccentricity of the generating ellipse of the ellipsoid . in section ( [ aplicacao ] ) we exhibit two simple applications of equations ( [ central ] ) . the case @xmath15 , despite being integrable , seems not to admit closed solutions . we show that , when @xmath16 , the only solutions for which the two vortices maintain their relative distance constant are the relative equilibria . for the @xmath17 case , we show numerically , by choosing specific parameters , that the three - vortices problem ( [ central ] ) on the ellipsoid is chaotic . in this section we discuss equations ( [ hally0 ] ) from a geometric and invariant point of view . hally s equations were deduced under the topological constraint that the sum of all vorticities is zero and , in fact , this is essential in recasting the equations in hamiltonian form . we also show , that , under this zero vorticity constraint , hally s equations on the sphere , coincide with bogolmonov s equations . let @xmath2 be a 2-dimensional riemannian manifold with metric tensor @xmath18 . @xmath2 is a symplectic manifold with symplectic form @xmath19 given by the area form induced by @xmath18 . let @xmath20 be the hamiltonian function . equations @xcite are @xmath21 where @xmath22 is the hamiltonian vector field induced by @xmath23 . now we assume that @xmath2 is conformal to the plane , that is , that there are coordinates @xmath24 where the metric is given by @xmath25 for some function @xmath26 . in these coordinates the symplectic ( area ) form is given by @xmath27 introduce complex coordinates @xmath28 on the plane @xmath29 by @xmath30 using @xmath28 we obtain @xmath31 and @xmath32 hamilton s equations ( [ hamil ] ) become @xmath33 for convenience we think of @xmath34 as the new area form and @xmath35 as the new hamiltonian function . writing @xmath36 the equations become @xmath37 we now show that hally s equations can be written as a hamiltonian system on the conformal plane . the equations are @xmath38 multiplying each term of the sum in the right hand side by @xmath39 and adding the zero @xmath40 to the right - hand side we obtain @xmath41 assuming that the sum of the vorticities is equal to zero we have @xmath42 that is @xmath43 but those are hamilton s equations @xmath44 for the symplectic two - form @xmath45 and hamiltonian function @xmath46 denote by @xmath47 the two - dimensional sphere @xmath48 @xmath49 is a riemannian manifold with the metric induced by the euclidean metric in @xmath50 . introduce spherical coordinates on @xmath49 trough the relations @xmath51 with @xmath52 , @xmath53 consider the conformal map @xmath54 , @xmath55 where @xmath56 the conformal factor is given by @xmath57 where @xmath58 now let @xmath59 . the euclidean distance @xmath60 in @xmath50 can be computed in the conformal variables . in fact @xmath61 that is @xmath62 now consider the map @xmath63 given by @xmath64 the pull back by @xmath65 of the symplectic form @xmath66 is just @xmath67 where by @xmath68 we denote the standard area form @xmath69 induced by the euclidean metric on @xmath70 . the pull back of the hamiltonian function @xmath71 is @xmath72 hamilton s equations for hamiltonian function ( [ hamm ] ) and symplectic form ( [ simp ] ) are precisely bogolmonov s equations for @xmath73 interacting vortices on a sphere . we notice that hally s equations are equivalent to bogolmonov s equations only for the case where all the vorticities sums up to zero . therefore , in dealing with the general vorticity case , one must use bogomolnov s equations . let @xmath74 represent the two dimensional ellipsoid of revolution @xmath75 where @xmath76 and @xmath77 if @xmath78 the ellipsoid is prolate . in this case the eccentricity @xmath79 of the generating ellipse is given by @xmath80 . if @xmath81 the ellipsoid is oblate and the eccentricity of the generating ellipse is given by @xmath82 we observe that @xmath83 and therefore the truncated first order system is expected to approximate the real system for quite high values of the eccentricity . we introduce coordinates on @xmath84 trough the relations @xmath85 with @xmath52 , @xmath53 if @xmath86 the surface is a sphere of radius @xmath87 . in this case the projection @xmath88 is a conformal transformation . we want to find a conformal transformation from @xmath84 to the conformal plane when @xmath89 . by the symmetry of the ellipsoid the conformal factor must depend only on the radial variable @xmath90 and not on the angular variable @xmath91 we look for a transformation of the form @xmath92 we observe that when @xmath86 the inverse of the above transformation is just transformation ( [ stereo ] ) written in polar coordinates . the symmetry of the ellipsoid implies that the equator is a circle of radius @xmath87 . in the @xmath93 coordinates the equator has parametrization given by @xmath94 and @xmath95 . therefore in the conformal coordinates @xmath96 the equator becomes the circle of radius @xmath97 we remark that the conformal factor for the ellipsoid , when restricted to the circle of radius @xmath98 is just the usual one for the sphere since the ellipsoid @xmath84 and the @xmath49 coincide at the equator . this is of utmost importance for what follows . in fact , since we will impose conformality at each order this trivial observation implies that @xmath99 for all @xmath3 . this is the condition that will guarantee the uniqueness of our series expansion . [ unique ] alternatively ( and equivalently ) , the uniqueness of the expansion can be obtained by imposing that the symmetry condition @xmath100 has to be satisfied at all orders . the algorithmic determination of the @xmath101 is obtained by imposing conformality at each order and follows the standard procedures of classical perturbation theory : truncate the series at order @xmath102 , find the conditions to be satisfied , and them assume those conditions to find the term of order @xmath103 . to find @xmath104 we impose that @xmath105 generates a conformal transformation up to first order in @xmath9 . the line element for the ellipsoid is @xmath106 we want to write @xmath107 in terms of @xmath90 and @xmath108 . to this goal it suffices writing @xmath109 and @xmath110 in terms of @xmath90 . but @xmath111 this gives @xmath112 also @xmath113 therefore @xmath114 finally @xmath115 \ , r^2 d \psi^2 + \right . \\ \\ + \left . \left [ 1 + \left ( \frac{4 \ , r^2}{(1+r^2)^2 } + 2 \ , ( 1 + r^2 ) \ , \frac{\p f_1(r)}{\p r } \right ) \ , \epsilon \ , + \mathcal o(\epsilon^2 ) \right ] \ , dr^2 \,\right\ } . \end{array}\ ] ] this will be conformal to the plane if it is a multiple of @xmath116 . we observe that if @xmath86 then the transformation is conformal with conformal factor @xmath117 . for @xmath118 the only way this transformation can be conformal for all values of @xmath9 is if the coefficient of @xmath9 in the first square bracket is equal to the coefficient of @xmath9 in the second square bracket , that is if @xmath119 now by remark ( [ unique ] ) we must have @xmath120 what implies @xmath121 . finally we have @xmath122 transformation ( [ trans ] ) becomes @xmath123 the conformal factor is , to first order given by , @xmath124 that is @xmath125 we write @xmath126 where @xmath127 for future reference we observe that in terms of the variables @xmath128 and @xmath108 we obtain @xmath129 the above expansion can not be carried explicitly further . in fact , the equation for the second order factor will be @xmath130 the general solution for this equation is given by @xmath131 where @xmath132 is an arbitrary constant . the above integral can not be computed using elementary functions . using the conformal factor just found we will write the n - vortex equations on the ellipsoid of revolution up to first order terms in @xmath9 . we use the coordinates @xmath93 introduced trough equations ( [ coord11 ] ) . the symplectic two - form becomes @xmath133 observing that @xmath134 we define the vector field @xmath135 . @xmath136 satisfies hamilton s equations @xmath137 and the flow @xmath138 of the vector field @xmath139 is a time reparametrization of the flow @xmath140induced by @xmath141 . in fact @xmath142 we therefore consider equation ( [ ham ] ) with @xmath23 and @xmath26 given by ( [ hamilconfor ] ) and ( [ expan ] ) respectively . @xmath23 is given by @xmath143 + \mathcal o ( \epsilon^2)\ ] ] the hamiltonian function @xmath23 is a function on the variables @xmath144 . moreover , the symplectic two - form ( [ ham ] ) is the standard symplectic form over the sphere . therefore , we write @xmath145 + \mathcal o ( \epsilon^2)\ , . \end{array}\ ] ] we conclude that , to first order in @xmath9 the dynamics of the n vortices on the ellipsoid is determined by the hamiltonian system with hamiltonian function @xmath146 \end{array}\ ] ] and symplectic two - form @xmath147 alternatively , in cartesian coordinates the above hamiltonian system becomes @xmath148 with symplectic two - form given by @xmath149 when @xmath15 the hamiltonian system given by ( [ hamn ] ) and ( [ simpn ] ) reduces to @xmath150 doing the following change of variables @xmath151 we obtain @xmath152 where @xmath153 @xmath154 equations ( [ mot2 ] ) are singular when @xmath155 i.e. @xmath156 . when @xmath86 equations ( [ mot2 ] ) reduces to a linear system and can be easily solved . those are the well known solutions for the two - vortex problem over the sphere : in a system of coordinates such that its z - axis is in the same direction as the center of vorticity @xmath157 the solutions are given in the @xmath158 , @xmath159 variables by @xmath160 where @xmath161 with @xmath108 a constant and @xmath162 with @xmath163 we observe that all solutions are periodic and the relative distance between the vortices is a first integral . when @xmath164 the center of vorticity vector @xmath157 is not preserved . in fact , only it s @xmath165 component @xmath166 is preserved . this first integral implies , by dimensionality that system is integrable . since @xmath167 we obtain that @xmath168 are constants of motion . despite being integrable system ( [ mot2 ] ) seems not to admit explicit solutions for a general initial condition . their relative equilibria can be easily characterized in the cases @xmath169 . then we obtain that @xmath170 first we characterize the solutions such that @xmath171 is constant along the motion . since @xmath172 and since for@xmath173 either @xmath174 or @xmath175 it follows that @xmath176 and @xmath177 are constants for the relative equilibria . therefore @xmath178 @xmath179 implies that @xmath180 is a constant and therefore @xmath181 . from the equations it follows that @xmath182 this implies that the only way ( [ ref1 ] ) can be zero is if @xmath183 we arrive at the result that in a relative equilibria @xmath176 , @xmath184 and @xmath180 are constants and that the vectors @xmath185 and @xmath186 are always in the same plane . the vectors @xmath157 and @xmath187 will precess around the @xmath165-axis with constant angular velocity and keeping their relative orientation . it follows from the definition that @xmath158 and @xmath159 will also precess around the @xmath165-axis with constant angular velocity and keeping their relative orientation . therefore the only solutions for which @xmath171 are constants are precisely the relative equilibria . case @xmath189 : if @xmath190 then @xmath191 and @xmath192 assuming that @xmath193 , @xmath194 , @xmath195 and @xmath196 we obtain that the angular velocity is given by @xmath197 we observe that the perturbation factor in the angular velocity is proportional to @xmath198 and therefore the deviation from the corresponding spherical relative equilibria should be bigger for smaller vorticities . we also observe that when @xmath86 then , taking as before @xmath199 : if @xmath200 then @xmath201 and @xmath202 proceeding as before we obtain @xmath203 the perturbation term is equal in both cases and again , as @xmath204 we will have @xmath205 the symmetry breaking when @xmath164 is expected to destroy the integrability of the three - vortices spherical problem . in fact , when @xmath206 the center of vorticity vector @xmath207 is conserved . the integrability follows by dimensionality . when @xmath118 , only it s z - component @xmath208 is conserved . that the integrability is destroyed is shown numerically in this section . we do not pursue a complete numerical study of the problem . this will addressed in a future work . to calculate the poincar sections for the three - vortex problem we will introduce an appropriate coordinate system . first we introduce a cylindrical system of coordinates @xmath209 with @xmath210 and @xmath211 for @xmath212 . the transformation is defined trough @xmath213 since @xmath214 this is a symplectic transformation and the equations in cylindrical coordinates are given by @xmath215 to reduce the symmetry we define the coordinates @xmath216 it follows that @xmath217 therefore the new system of coordinates forms a canonical system . we observe that @xmath218 and therefore @xmath219 . then @xmath220 is a cyclic variable and the hamiltonian ( [ hamn ] ) is given by @xmath221 since this hamiltonian is now defined in a four dimensional phase - space , the poincar maps can be visualized . the equations are integrated numerically using the fehlberg-7(8 ) runge - kuta . the relative error in energy is kept smaller then @xmath222 . a constant step size equals to @xmath223 is used in all integrations . figure 1 show the pictures of three poincar maps . the onset of chaotic behavior is clear . it appears ( as expected ) around the homoclinic solution present in the integrable case @xmath224 . we observe that for @xmath225 and @xmath226 the eccentricity of the generating ellipses are @xmath227 and @xmath228 respectively . marsden , j. e. ; ratiu , t. s. introduction to mechanics and symmetry . a basic exposition of classical mechanical systems . second edition . _ texts in applied mathematics , 17 . _ springer - verlag , new york , 1999 .

we consider the n - vortex problem on a ellipsoid of revolution . applying standard techniques of classical perturbation theory we construct a sequence of conformal transformations from the ellipsoid into the complex plane . using these transformations the equations of motion for the n - vortex problem on the ellipsoid are written as a formal series on the eccentricity of the ellipsoid s generating ellipse . first order equations are obtained explicitly . we show numerically that the truncated first order system for the three - vortices system on the symmetric ellipsoid is non - integrable . @xmath0departamento de matemtica universidade federal de pernambuco recife , pe cep 50740 - 540 brazil and @xmath1 the abdus salam ictp strada costiera 11 trieste 34100 italy 1.0 cm * key words : * vortex dynamics , hamiltonian systems , perturbation theory . 1.0 cm * msc : * 76c05 , 58f40 , 70h05 .

star formation occurs inside giant molecular clouds ( gmcs ) and surveys of molecular gas in galaxies show that it is typically concentrated in large complexes or spiral arm segments having sizes up to a kiloparsec and masses up to @xmath0 @xmath1 ( solomon & sanders,1985 and elmegreen , 1993 ) . these complexes may contain several giant molecular clouds ( gmcs ) with sizes up to @xmath2 parsecs ( pc ) and masses up to @xmath3 . the gmcs in turn , contain much smaller scale structures that may be filamentary or clumpy on a wide range of scales ( blitz , 1993 ; blitz & williams , 1999 ; williams , blitz & mckee , 2000 ) . the substructures found in gmcs range from massive clumps with sizes of several parsecs and thousands of solar masses , which may form entire clusters of stars , to small dense cloud cores with sizes of the order of @xmath4 pc and masses of the order of @xmath5 ( mendoza , 1985 ; cernicharo,1991 ; larson , 1994 ; williams , blitz & mckee , 2000 ; andre , ward - thompson & barsony , 2000 ; visser , richer & chandler , 2002 ) . the internal structure of molecular clouds is partly hierarchical , consisting of smaller subunits within larger ones ( scalo,1990 ; larson,1995 ; simon , 1997 ; stutzki et al . , 1998 ; elmegreen , 2000 ) . in particular , the irregular boundaries of molecular clouds have fractal - like shapes resembling those of surfaces in turbulent flows , and this might mean that the shapes of molecular clouds are created by turbulence ( falgarone , phillips & walker , 1991 ; falgarone , puget & perault , 1992 ) . + most molecular clouds form stars , but very inefficiently , typically turning only a few percent of their mass into stars before being dispersed . despite the strong dominance of gravity over thermal pressure , this low efficiency has long been considered problematic , and implying that additional effects such as magnetic fields , angular momentum conservation or turbulence support these clouds in near - equilibrium against gravity and prevent a rapid collapse ( heiles et al . , 1993 ; mckee et al . , 1993 ) . + the lifetimes and evolution of molecular clouds is provided by the ages of the associated newly formed stars and star clusters ( blaauw , 1991 ; larson,1994 ; elmegreen , 2000 ; andre , ward - thompson & barsony , 2000 ; hartmann , ballesteros - paredes & bergin , 2001 ) . very few gmcs are known that are not forming stars , and the most massive and dense ones as a rule contain newly formed stars . they also can not survive for long after beginning to make stars , since the age span of the associated young stars and clusters is never more than @xmath6 myr , about the dynamical or crossing time of a large gmc . stars and clusters older than @xmath7 myr do not appear to be associated with molecular gas ( leisawitz , bash & thaddeus , 1989 ) . + the youngest stars are associated with the denser parts of molecular clouds , and especially with the densest cloud cores that appear to be the direct progenitors of stars and stellar groupings ( mendoza,1985 ; lada , strom & myers , 1993 ; williams , blitz & mckee , 2000 ; andre , ward - thompson & barsony , 2000 ) . + after several decades of study , the physical conditions in nearby star - forming molecular clouds are now fairly well understood , at least for the smaller nearby dark clouds that form mostly low - mass stars . however in the past @xmath8 years , infrared observations have revolutionized our understanding of star formation ( gutermuth et al . , 2009 ) . the emission excess of young stars is well above that expected from reddened stellar photospheres and originates from the dusty circumstellar disks and envelopes surrounding young stars . for these reasons , colour - colour diagrams and cmds in the infrared ( ir ) have proven to be excellent tools for identifying and classifying young stellar objects ( megeath et al . 2004 ; harvey et al . 2008 ; gutermuth et al . + we intend to address the physical mechanisms responsible for the formation of pms stars in embedded star clusters which is crucial for understanding their properties . + in sect . [ history ] we discuss the ic5146 nebular complex to which a streamer appears to be related , and contains a sample of 5 newly found embedded clusters ( ecs ) . in sect . [ sample ] we gather the data . in sect . [ cmds ] we discuss the general properties of the objects . in sect . [ disc ] we discuss the results . finally , in sect . [ conclu ] we provide the conclusion of this study . ic5146 is a reflection and emission nebula in the cygnus constellation centred at about @xmath9 , @xmath10 . to the east of the cloud there occurs a streamer . the streamer region corresponds to a long dark nebula discussed by wolf ( 1904 ) . hubble ( 1922 ) classified it in the optical as continuous , whereas minkowski ( 1947 ) listed the nebula to have h@xmath11 emissions . both are right , since ic5146 is a transition case between a reflection nebula and an hii region . many surveys at different wavelengths have been carried out in this region and associated clouds ( e.g. kramer et al . , 2003 ) . [ ic5146 ] shows an optical image of the ic5146 nebular and the streamer ( @xmath12 ) extending to the east . in the present study , one of us ( e.b . ) discovered 5 new ecs in the area of ic5146 using wise ( fig . [ figxx ] and [ figxxx ] ) . the w2 images in fig . [ figxx ] enhance the stellar and protostellar components of the clusters , while a w4 is sensitive to extended dust emission . sources brighter in w4 than in w2 are certainly due to the presence of dust envelopes or disks . nbb5 is projected close to ic5146 and all the clusters candidates apperar to be immersed in cold molecular material . the central coordinates of these ecs are given in table [ tab1 ] , together with the estimated angular radii . although the cluster stellar concentrations may not be evident in figs . [ figxx ] and [ figxxx ] , the embedded clusters , in general , have conspicuous radial density distributions in fig . [ rdps ] . the ic5146 region has an important population of pms stars . embedded sources , sometimes known as protostars are optically invisible young stars . however , they are observed in the infrared and the majority is associated with the cloud ( see sect . 3).we will explore ic5146 and the 5 ecs in the nebular complex using colour - colour diagrams and cmds ( bonatto & bica , 2007 ; camargo , bica & bonatto , 2015 ) . in this work , ic5146 refers either to the nebula or the cluster . .the embedded cluster ic 5146 and the 5 newly found ecs along the streamer [ cols="<,^,^,^,^,^,^,^ " , ] table notes - the adopted methods and references are : johnson & hilter ( 1956)-zams ; [ 2 ] walborn ( 1972)-zams ; [ 3]blaauw ( 1963)-zams ; [ 4 ] balona & feast ( 1975)-zams ; [ 5 ] star counts ; [ 6 ] allen et al.(1982)-zams ; [ 7 ] jaschek & gomez ( 1998)-ms ; [ 8 ] harvey et al.(2008)-onc zams ; [ 9 ] pms and ms . the majority of the embedded clusters do not survive the first few ten myr , especially the low - mass ones , given the interplay among environmental conditions , star - formation efficiency , initial stellar dynamical state and mass fraction converted into stars or expelled . it is important to investigate the structural and photometric properties in this early phase . we report the discovery of @xmath13 ecs along the streamer to the east of ic5146 and compare their cmds and colour - colour diagrams , suggesting a similarity of properties among them . the features in the colour - colour diagrams of the streamer and the 5 ecs are compatible with an age of @xmath14 @xmath13 myr ( fig . [ ccs ] and table [ tab2 ] ) . in principle , the detailed distribution of stars within the pms area in the cmd plane reflects the star formation history of the embedded population . the relative mean ages of young clusters can be established to greater precision by using a single or a consistent set of pms models to extract the cluster ages ( e.g. , haisch et al . , 2001a and bressan et al . , 2012 ) in all five new ecs ( table [ tab2 ] ) , the young stars identified by their excess emission in the mid - ir are possibly distributed in a range of distances . in contrast , the diameters of clusters by star counts are typically 4 pc or less ( lada & lada , 2003 ) . this confirms that a significant fraction of gas in each star - forming region forms outside and surrounding the clustered region . the distances and uncertainties suggest that the nbbs clusters are associated physically . possibly nbb2 is a bit far . the 2mass cmds ( fig.[cmds ] ) and estimated number of stars in column 7 in table [ tab1 ] provide that ic5146 is by far , the most populous in the sample . futhermore , we suggest that star formation in ic5146 is continuing in the halo , while the molecular gas has been dispersed toward the cluster core and streamer . megeath et al . 2004 & gutermuth et al . 2004 have discussed yso classifications . [ fig_8 ] shows matched irac colors of ic5146 , together with model loci by allen et al . the data seem to cluster into three main regions : ( i ) a clump around @xmath15 $ ] that contains mostly background / foreground stars and class iii sources with no intrinsic ir excess , ( ii ) a clump that occupies the class ii region ( within the large red box ) , and a group that runs along the class i locus . sources that lie between the @xmath16 colors of stellar photospheres and class iii stars @xmath17 $ ] and class ii sources @xmath18 $ ] can be understood if we suggests the effects of dust grain growth and settling . in addition , there are a few sources that do not lie inside the class ii locus but do lie along the extinction vector and could therefore be reddened class ii objects . the most extreme points might be trace filaments and high opacity ( dark even at @xmath19 ) , consistent with the large @xmath20 implied by this scenario . alternatively , they could be extremely low luminosity class i sources . thus , the class i stage ( fig . [ cmds ] to [ fig_9 ] ) is a later stage of protostellar collapse . the class ii stage ( fig . [ fig_8 ] ) is characterized by the presence of excess of ir emission above that expected for a stellar photosphere . these stars can be inside it or lie along the extinction vector and could therefore be reddened . this is the signature of a classical t tauri surrounded by an accretion disk . infall from the cloud has ceased owing to dispersal of the remnant infall envelope by the combined effects of infall and outflow ( allen et al . 2004 ; megeath et al . 2004 ; sung , bessell & lee , 1997 ) . the class iii stage can last for about @xmath0yr ending in a zero - age main - sequence star ( palla , 1999 ) . the distribution of sources in each ec is strikingly similar , but some distinct concentrations of stars are also apparent ( figs . [ cmds ] to [ fig_str ] ) . all ecs in the sample have central density excesses ( fig . [ rdps ] ) . the ecs appear to have an extra dust content , possibly related to a younger age . on the other hand , ic 5146 has a prominent rdp core . the rdps in fig . 5 were constructed with spitzer data , except for nbb1 that is not in the spitzer field , so its rdp was built with wise data . ages throughout the complex are approximately constant at @xmath21 myr ( table [ tab2 ] ) . v absorptions are strong . ic5146 and the streamer characterize the complex , at distance @xmath22 kpc , nbb2 , nbb3 and nbb4 appear to lie a bit farther ( table [ tab2 ] ) . the streamer is an optical dark nebula ( fig.1 ) , that in the near ir is composed of many stars . the streamer is a prestellar region ( roy et al . 2011 , kramer et al . 2003 ) with essencially no emission from warm grains . it might be a particular stage in evolution of clouds , with a mix dust phases at different temperatures . dobashi et al . ( 2003 ) have suggested that outflows have played an important role in supporting the parent cloud from collapsing . we conclude that the streamer is stiil a _ box to be further explored . the present analyses suggest that the ic5146 cluster , the streamer and the 5 newly found embedded clusters are neighbouring primeval examples of extended star formation and they are crucial for future more detailed studies of ic5146 complex , and * for comparisons with other star forming regions . * we thank paul harvey , the university of texas at austin and the go-@xmath23 spitzer legacy project for providing us with spitzer data . we make use of the 2mass and wise databases . we thank an anonymous referee for important remarks . we acknowledge financial support from the brazilian institution cnpq . 00 blaauw , a. 1991 , the physics of star formation and early stellar evolution , nato advanced science institutes ( asi ) series c , vol . 342 , held in agia pelagia , crete , greece , may 27th - june 8th , dordrecht : kluwer , 1991 , edited by charles j. lada and nikolaos d. kylafis . , p.125 cernicharo , j. 1991 , the physics of star formation and early stellar evolution , nato advanced science institutes ( asi ) series c , vol . 342 , held in agia pelagia , crete , greece , may 27th - june 8th , dordrecht : kluwer , 1991 , edited by charles j. lada and nikolaos d. kylafis .

throughout the last decade sensitive infrared observations obtained by the spitzer space telescope significantly increased the known population of ysos associated with nearby molecular clouds . with such a census recent studies have characterized pre - main sequence stars ( pms ) and determined parameters from different wavelengths . given the restricted spitzer coverage of some of these clouds , relative to their extended regions , these yso populations may represent a limited view of star formation in these regions . we are taking advantage of mid - infrared observations from the nasa wide field infrared survey explorer ( wise ) , which provides an all sky view and therefore full coverage of the nearby clouds , to assess the degree to which their currently known yso population may be representative of a more complete population . we extend the well established classification method of the spitzer legacy teams to archived wise observations . we have adopted 2mass photometry as a `` standard catalogue '' for comparisons . besides the massive embedded cluster ic5146 we provide a multiband view of five new embedded clusters in its surroundings that we discovered with wise . in short , the analysis involves the following for the presently studied cluster sample : _ ( i ) _ extraction of 2mass / wise / spitzer photometry in a wide circular region ; _ ( ii ) _ field - star decontamination to enhance the intrinsic colour - magnitude diagram ( cmd ) morphology ( essential for a proper derivation of reddening , age , and distance from the sun ) ; and ( iii ) construction of colour - magnitude filters , for more contrasted stellar radial density profiles ( rdps ) . galaxy , molecular cloud , yso , embedded clusters : general , photometry

since the first discoveries @xcite , the number of known extra - solar planets has been rapidly increasing and it reaches @xmath2 ( http://exoplanet.eu : * ? ? ? * ) . a significant fraction of these planets are members of multi - planet systems @xcite , which were mostly discovered by the radial - velocity and transit methods . multi - planet systems can also be detected by using the microlensing method . gravitational microlensing occurs by the chance alignment along the line of sight toward a background star ( source ) and a foreground object ( lens ) , resulting in the magnification of the source brightness due to the bending of light by the gravity of the lens . when a lensing object is a star accompanied by a planet , the planet causes astigmatism in the bending of light and induces formation of caustics which refers to the envelope of light rays on the source plane at which the magnification of a point source becomes infinity . a planetary lensing signal occurs when the source approaches caustics and it appears as a short term anomaly superposed on the smooth and symmetric lensing light curve of the host star @xcite . if a planetary system is composed of multiple planets , each planet induces its own caustics and the multiplicity of the planets can be identified if the source trajectory passes multiple caustics induced by the individual planets @xcite . the sensitivity to multiple planets for general microlensing events is low because planet - induced caustics are small and thus the geometrical probability for the source to pass multiple caustics is low . however , the sensitivity is high for a subset of lensing events where the brightness of a source star is greatly magnified @xcite . this is because one of the caustics induced by each planet is always located very close to the host star around which the source of a high - magnification event passes by @xcite . indeed , two multi - planet systems identified by microlensing @xcite were discovered through this channel . microlensing discoveries of multi - planet systems are important for better understanding of planet - formation mechanism . according to the core accretion theory @xcite , giant planets are believed to form in group in the region beyond the snow line where the lower temperature in this region makes many more solid grains available for accretion into planetesimals . however , most known planets in multiple systems reside well within the snow line . these close - in planets are believed to have migrated from the outer region where they were formed @xcite . detecting planets at their birth places by using the transit and the radial - velocity methods is difficult due to their low sensitivity to wide - separation planets . by contrast , microlensing is sensitive to planets located around and beyond the snow line and thus provides a channel to study planets where they formed . in this work , we reanalyze the light curves of the eight high - magnification microlensing events for each of which a single planet was previously detected . from the analyses , we investigate the existence of additional planets . lrrrl ogle-2005-blg-071 & 59 & 3.8@xmath3 @xmath4 & 3.6@xmath5 & @xcite , @xcite + ogle-2005-blg-169 & 880 & 13@xmath6 @xmath7 & 2.7@xmath8 & @xcite + moa-2007-blg-400 & 629 & 0.83@xmath9 @xmath4 & 0.72@xmath10 / 6.5@xmath11 & @xcite + moa-2008-blg-310 & 366 & 74@xmath12 @xmath7 & 1.25@xmath13 & @xcite + moa-2009-blg-319 & 208 & 50@xmath14 @xmath7 & 2.4@xmath15 & @xcite + moa-2009-blg-387 & 49 & 2.6@xmath16 @xmath4 & 1.8@xmath17 & @xcite + moa-2010-blg-477 & 599 & 1.5@xmath18 @xmath4 & 2.0@xmath19 & @xcite + moa-2011-blg-293 & 418 & 2.4@xmath20 @xmath4 & 1.0@xmath21 / 3.4@xmath3 & @xcite in published articles , there exist 29 planetary microlensing events from which 31 planets were found . among them , events for our analyses are selected based on the following criterion . first , we choose planetary events with high magnifications due to their high sensitivity to multiple planets . with an adopted threshold magnification @xmath22 , there exist 15 events . among the planetary systems found from these events , two are already known to be in multi - planet systems @xmath23 is orbited by two planets with masses @xmath24 @xmath4 and @xmath25 @xmath4 and separations from the host @xmath26 au and @xmath27 au , respectively @xcite . the second system is ogle-2012-blg-0026l , where a host star with a mass @xmath28 @xmath23 is orbited by two planets with masses @xmath29 @xmath4 and @xmath30 @xmath4 and separations @xmath31 au and @xmath27 au , respectively @xcite . ] and another two are in stellar binary systems . @xmath7 ) lies at @xmath32 au from its host with a mass @xmath330.100.15 @xmath23 and the host itself orbits another star with a mass @xmath34 @xmath23 and a separation @xmath331015 au @xcite . the second microlensing planet in binary is ogle-2008-blg-092 lb , where a @xmath35 @xmath7 planet orbits a @xmath36 @xmath23 star at @xmath37 au . the host has a low - mass companion with @xmath34 @xmath23 @xcite . ] we exclude the events with already - known multi - planet systems and circum - stellar binary planets from our analysis . second , we choose events with good coverage of the peak region . this is because signatures of multiple planets for high - magnification events show up near the peak of a light curve and thus good coverage of the peak region is essential to constrain additional planets . third , we confine analysis to events for which there is no ambiguity in the interpretation of the known planetary signals . applying these criteria leave 8 events including ogle-2005-blg-071 , ogle-2005-blg-169 , moa-2007-blg-400 , moa-2008-blg-310 , moa-2009-blg-387 , moa-2010-blg-477 , and moa-2011-blg-293 . in table [ table : one ] , we list the lensing events that we analyzed . also presented are the peak magnifications @xmath38 of the individual events , the physical parameters ( mass and separation from the host ) of the previously discovered planets , and the references of the original analyses . in figure [ fig : one ] , we present the light curves of the events . we note that the same data sets as those used in the original analysis are used in our analysis for the consistency of results . in order to investigate the existence of an additional planet for a lensing event with an already known planet , it is required to conduct triple - lens modeling . for a triple - lens system , the image positions for given locations of the lens and source are found by solving the equation of lens mapping ( lens equation ) that is expressed as @xmath39 here @xmath401 , 2 , and 3 denote the individual lens components , @xmath41 , @xmath42 and @xmath43 are the complex notations of the source , lens , and image positions , respectively , @xmath44 denotes the complex conjugate of @xmath45 , and @xmath46 is the mass fraction of each lens component @xcite . once the image locations are found , the lensing magnification is computed by @xmath47 where @xmath48 is the magnification of each image . due to non - linearity , the lens equation for a triple lens system can not be analytically solved . however , lensing magnifications can be numerically computed by using the method known as the inverse ray - shooting method @xcite . in this method , a large number of rays are uniformly shot from the image plane , bent by the lens equation , and land on the source plane . then , lensing magnifications are computed as the ratio of the number density of rays on the source plane to that on the image plane . the lens equation can also be solved semi - analytically because the equation is expressed as a tenth - order polynomial in @xmath45 and the image positions and subsequent lensing magnifications are computed by solving the polynomial @xcite . the semi - analytic method has an advantage of fast computation but it is difficult to be used for computing lensing magnifications affected by finite - source effects . on the other hand , the numerical method can be used to compute finite - source magnifications but computation is slow . describing the light curve of a triple - lensing event requires ten basic parameters . the first three of these parameters describe how the source approaches the lens . they are the time of the closest source approach to a reference position of the lens system , @xmath49 , the normalized projected separation between the source and the lens reference position at that moment , @xmath50 , and the einstein time scale , @xmath51 , that is defined as the time required for the source to cross the angular einstein radius of the lens @xmath52 . another six parameters characterize the lens system . they are the projected separations between the host and the individual planet companions , @xmath53 and @xmath54 , and the mass ratios of the companions to the host , @xmath55 and @xmath56 , the angle between the source trajectory and the axis connecting the host and the first planet , @xmath57 , and the orientation angle of the second planet with respect to the host - first planet axis , @xmath58 . we note that length scales describing lensing phenomenon are usually expressed in units of @xmath52 and @xmath50 , @xmath53 and @xmath54 are normalized by @xmath52 . the last parameter is the angular size of the source @xmath59 in unit of the einstein radius , @xmath60 ( normalized source radius ) . this parameter is needed to describe light curve deviations affected by finite - source effects . besides the basic lensing parameters , precise description of lensing light curves often requires additional parameters to account for higher - order effects . one such effect is caused by the positional change of an observer induced by the orbital motion of the earth around the sun . considering this parallax effect @xcite requires two additional parameters @xmath61 and @xmath62 that represent the two components of the lens parallax vector @xmath63 projected on the sky in the north and east equatorial coordinates , respectively . similarly , lensing light curves can also be affected by the change of the lens position caused by the orbital motion @xcite . with the lensing parameters , we conduct triple - lens modeling of the events that were previously analyzed based on binary - lens modeling . for all analyzed events , the anomaly is dominated by the signal of the already reported planet and thus the signal of the potential additional planet would be small . we , therefore , set the lensing parameters related to the known planet @xmath64 fixed as outlined in @xcite . the existence of an additional planet is then inspected by checking whether the two - planet model improves the fit with respect to the single - planet model . not knowing the characteristics of the second planet , we search for the second - planet parameters @xmath54 and @xmath56 by inspecting solutions in wide ranges of the parameter space spanning @xmath65 and @xmath66 , respectively . for a given set of @xmath67 , the other lensing parameters are searched for by minimizing @xmath68 in the parameter space using markov chain monte carlo ( mcmc ) algorithm . for computing triple - lensing magnifications , we apply both the semi - analytic and numerical methods . for the region near the peak of the light curve , we use the numerical inverse ray - shooting method because this region is likely to be affected by finite - source effects during the source star s approach close to the planet - induced caustics . for other regions , we use the semi - analytic method for fast computation . computing finite - source magnifications is further accelerated by applying the `` map making '' method @xcite . in this method , a map of rays for a given set of @xmath64 and @xmath67 is constructed based on the inverse ray shooting method and then it is used to produce many different light curves resulting from various source trajectories without producing extra maps . in computing finite - source magnifications , we consider limb - darkening variation of the source star surface brightness by modeling the surface brightness profile as @xmath69 , where @xmath70 is the limb - darkening coefficients corresponding to specific passband @xmath71 and @xmath72 is the angle between the line of sight toward the source star and the normal to the source surface . we use the same limb - darkening coefficients that were used in the previous analyses where first planets were reported . if higher - order effects were reported in the previous analyses , we also consider them . parallax effects were reported for events ogle-2005-blg-071 , moa-2009-blg-387 , and moa-2010-blg-477 . on the other hand , obvious effect of lens - orbital motion was reported for none of the events . we note that the effect of an additional planet is confined to the narrow region of the light curve peak while both parallax and lens - orbital effects influence the overall shape of a lensing light curve . therefore , the existence of the second planet does not affect the lensing parameters of higher - order effects . table [ table : two ] shows the result of the analysis presented as @xmath68 values of the binary and triple lens models . from the comparison of the @xmath68 values , it is found that including an additional planet improves fits with @xmath0 for seven out of the eight analyzed events . for moa-2009-blg-319 , the improvement is relatively big with @xmath73 . lrrrccc ogle-2005-blg-071 & 1305.6 & 1283.7 & 21.9 & 1.5 5.1 & & + ogle-2005-blg-169 & 533.4 & 518.5 & 14.9 & 0.6 11.3 & 1.0 7.1 & + moa-2007-blg-400 & 770.8 & 760.0 & 10.8 & 0.8 4.5 & 1.2 2.6 & + moa-2008-blg-310 & 3154.4 & 3076.4 & 78.0 & 0.7 2.0 & 1.1 1.3 & + moa-2009-blg-319 & 6180.2 & 6037.5 & 142.7 & 0.5 7.9 & 0.9 5.0 & 1.6 2.9 + moa-2009-blg-387 & 1298.4 & 1269.3 & 29.1 & 0.9 4.4 & 1.6 2.5 & + moa-2010-blg-477 & 4444.0 & 4394.4 & 49.6 & 0.8 12.1 & 1.7 5.8 & + moa-2011-blg-293 & 5089.2 & 5069.9 & 19.3 & 1.0 3.5 & & although the improvement of the fit for the event moa-2009-blg-319 is formally significant , one can not exclude the possibility of systematics in data , which is often masqueraded as planetary signals . in order to check the genuineness of the additional planetary signal , we inspect the fits from the triple and binary lens models . figure [ fig : two ] shows the peak portion of the light curve . we plot data sets from moa , ctio , and ftn observations for which most signal of the additional planet comes from : @xmath74 for moa , @xmath75 for ctio , and @xmath76 for ftn data . also plotted are the light curves of the triple and binary models . it is found that the two models can not be distinguished within the resolution of the plot . to better show the difference between the two models , we also present the residuals from the individual models . we find that the improvement by the triple - lens fit is @xmath77 mag level which is equivalent to the scatter of the data , implying that the signal can centainly be attributed to the systematics in data . we find similar results for other events . therefore , we conclude that there is no case of firm detection of additional planets . although no clear evidence of additional planets is found , it is still possible to constrain the existence of additional planets around the lens . the constraint is expected to be strong because all analyzed events are highly magnified and thus efficiency to planets would be high in wide ranges of planet parameters . for the purpose of constraining additional planets , we construct _ exclusion diagrams _ @xcite showing the confidence levels excluding the existence of an additional planet as a function of its separation and mass ratio . to construct exclusion diagrams , we first estimate the detection efficiency of planets with respect to the planet parameters @xmath54 and @xmath56 according to the following procedure . 1 . we first choose a triple - lens configuration with @xmath78 . for a given set of these parameters , we determine the remaining parameters that best fit the observed light curve and then compute @xmath79 between the triple and binary fits . 2 . we repeat the above process for many different orientation angles @xmath58 of the second planet . the detection efficiency for a given @xmath80 is estimated as the fraction of the angles @xmath58 that produce significant deviations in the light curve . then , the confidence level of excluding an additional planet corresponds to the efficiency . as a criterion for planet detections , we set a threshold value @xmath81 . @xcite suggested plausible values in the range @xmath82 based on their experience in fitting lensing light curves for which planets were detected through the high - magnification channel . @xcite suggested a similar threshold . in our analyses , we adopt a median value @xmath83 . figure [ fig : three ] shows the exclusion diagrams of the individual events . for each diagram , the regions with different confidence levels of excluding additional planets are marked in different colors . the star mark corresponds to the planet detected in the previous analysis . the upper @xmath84 label represents the projected planet - host separation in a physical unit ( au ) . the ticks marked by `` j '' , `` s '' , `` u '' , and `` e '' represent the planet masses corresponding to the jupiter , saturn , uranus , and earth of the solar system , respectively . we note that the conversion from @xmath80 to the physical separation and mass of the planet is done based on the physical einstein radius @xmath85 and total mass of the lens system @xmath86 determined in the previous analyses , i.e. , @xmath87 and @xmath88m_{\rm tot}$ ] . in table [ table : two ] , we present the exclusion ranges of additional planets determinded with 90% confidence level for jupiter , saturn , and uranus - mass planets . it is found that the constraint on giant planets is considerably strong . on the other hand , the constraint on planets with masses less than uranus is relatively weak . the weak constraint on low - mass planets is due to the small size of the planet - induced central caustic which decreases rapidly with the decrease of the planet mass @xcite . in addition to planet masses , the sensitivity depends on various other factors . some of these factors are related to observation , including the photometric precision , cadence , and completeness of the planetary signal coverage , etc . other factors are related to the intrinsic properties of planetary events , including the peak magnifications , severeness of finite - source effects , etc . for ogle-2005-blg-169 , moa-2009-blg-319 , and moa-2010-blg-477 , the constraint is strong . these events have a common property of high peak magnifications with @xmath89 , @xmath90 , and @xmath91 , respectively . by contrast , the constraint is relatively weak for ogle-2005-blg-071 @xmath92 and moa-2009-blg-387 @xmath93 because the peak magnifications are low . for planets either with low masses on located away from the einstein ring of the host , the central caustic is small and thus signals of low - mass planets can be detected only for very high - magnification events where the source trajectories approach close to the caustic . for moa-2007-blg-400 and moa-2008-blg-310 , the peak magnifications are high : @xmath94 and @xmath95 , respectively . nevertheless , the sensitivity is relatively low . we find the main cause of the low sensitivity for these events is severe finite - source effects , which wash out planetary signals , especially for low - mass planets @xcite . for moa-2011-blg-293 , on the other hand , the sensitivity is not very high although the event was highly magnified @xmath96 and experienced little finite - source effects . we find that the low sensitivity is mainly due to low photometric precision caused by the source faintness . in order to investigate the potential existence of additional planets , we reanalyzed high - magnification microlensing events for each of which a single planet had been previously detected . we found that introducing additional planets improves fits but the levels of improvement are not big enough to firmly identify additional planets . although no clear evidence of additional planets was found , we could constrain the existence of additional planets in the parameter space . for this purpose , we presented exclusion diagrams showing the confidence levels excluding the existence of an additional planet as a function of its separation and mass ratio . we also presented the exclusion ranges of additional planets with 90% confidence level for jupiter , saturn , and uranus - mass planets .

high - magnification gravitational microlensing events provide an important channel of detecting planetary systems with multiple giants located at their birth places . in order to investigate the potential existence of additional planets , we reanalyze the light curves of the eight high - magnification microlensing events for each of which a single planet was previously detected . the analyzed events include ogle-2005-blg-071 , ogle-2005-blg-169 , moa-2007-blg-400 , moa-2008-blg-310 , moa-2009-blg-319 , moa-2009-blg-387 , moa-2010-blg-477 , and moa-2011-blg-293 . we find that including an additional planet improves fits with @xmath0 for seven out of eight analyzed events . for moa-2009-blg-319 , the improvement is relatively big with @xmath1 . from inspection of the fits , we find that the improvement of the fits is attributed to systematics in data . although no clear evidence of additional planets is found , it is still possible to constrain the existence of additional planets in the parameter space . for this purpose , we construct exclusion diagrams showing the confidence levels excluding the existence of an additional planet as a function of its separation and mass ratio . we also present the exclusion ranges of additional planets with 90% confidence level for jupiter , saturn , and uranus - mass planets .

the inverse compton scattering ( ics ) of relativistic electrons is one of the major radiation mechanisms in high energy astrophysics . since the compton cooling time of electrons decreases linearly with energy , this channel becomes especially prolific in the gamma - ray band . the universal presence of dense radiation fields makes ics an effective gamma - ray production mechanism in various astronomical environments , in particular in sources containing relativistic outflows - microquasars , agn , grbs , _ etc . _ the dramatically enhanced fluxes , the shift of the spectral energy distribution towards higher energies , and the reduction of characteristic timescales are distinct features of the doppler boosted radiation produced in a relativistically moving source with a doppler factor @xmath0 . in relativistic outflows , ics can be realized , depending on the origin of target photon fields , in two different scenarios . in the so - called synchrotron - self - compton ( ssc ) model the synchrotron photons of relativistic electrons constitute the main target for compton scattering of the same electrons . in its simplified homogeneous one - zone version , the ssc model assumes that a spherical source ( a `` blob '' ) filled with an isotropic electron population and a random magnetic field , moves towards the observer with a constant doppler factor @xmath1 . the beaming pattern for emission of a relativistically moving source derived by @xcite , can be directly applied to the synchrotron and ssc components of radiation . in the second scenario , the seed photon field for ics is dominated by external radiation , i.e. by low - energy photons produced outside the moving source of relativistic electrons . in this , the so - called external compton ( ec ) model the beaming pattern of the inverse compton radiation differs significantly from the beaming pattern of the synchrotron and ssc components @xcite . the characteristics of ec radiation of a relativistically moving source can be calculated using two different approaches : \(i ) transforming first the external radiation photon field to the jet frame , calculating in the same frame the characteristics of the high energy photon ( the outcome of ics ) , and finally transforming the latter back to the observer frame ; \(ii ) transforming the electron distribution from the jet frame to the observer frame , and then calculating the spectrum of ic photons directly in the observer frame . using the first approach , @xcite has derived the ec beaming pattern in the thomson limit assuming a power - law distribution of electrons . @xcite using the second approach have extended this result to the general case of compton scattering , including the klein - nishina regime . in both treatments , the distribution of relativistic electrons has been assumed to be isotropic and homogeneous . in the zeroth approximation , this could be a reasonable assumption , and thus can be in principle applied to the interpretation of gamma - ray observations of many blazers . however , in some other cases one can not exclude significant deviations of distributions of electrons from homogeneous and isotropic realizations . the non - stationary treatment of the problem is another issue which has not been addressed so far . in this paper , we develop a new approach which provides a strict formalism for the treatment of the beaming pattern for ec radiation . namely , we solve the photon transfer equation which allows us to the beaming pattern in a concise way , and , more importantly , to extend the formalism to more general ( non - stationary and anisotropic ) case of electron distributions , but assuming isotropic , homogeneous and isotropic distribution of the seed ( target ) photon fields . for demonstration of the potential of the proposed formalism , we examine how anisotropy in the electron distribution affects the the energy spectrum and the angular distribution of ec radiation . in two close points located on the line of sight , the specific intensity ( the spectral radiance ) @xmath2 and emissivity @xmath3 are related as ( see e.g. @xcite ) [ ec1 ] i(s)-i(s - ds)=jds , where @xmath4 is the distance along the line of sight . it is assumed than one can neglect the scattering and absorpion of radiation during its propagation . it is convenient to consider @xmath2 and @xmath3 as functions of the radius - vector @xmath5 , the time @xmath6 , and the wave - vector of the photon @xmath7 , i.e. @xmath8 , and @xmath7 , i.e. @xmath9 . for simplicity , hereafter we will use the system of units in which the speed of light @xmath10 and the planck constant @xmath11 . also , instead of @xmath2 and @xmath3 we will use the distribution function of photons @xmath12 and the source function @xmath13 , where @xmath14 is the photon energy . the function @xmath15 describes the number of photons in the volume element @xmath16 of the phase space at the moment @xmath6 , while @xmath17 is the number of photons in the momentum interval @xmath18 , emitted during the time interval @xmath19 from the volume @xmath20 of the source located at the point @xmath5 . with these new notations , eq . ( [ ec1 ] ) can be written in the form [ ec2 ] g(k , r , t)-g(k , r - n ds , t - ds)=q(k , r , t)ds , where @xmath21 is a unit vector along the photon momentum . using eq . ( [ ec2 ] ) , it is easy to derive the following relation [ ec3 ] g(,,t)=^_0 ds q(,r - ns , t - s ) , which expresses @xmath22 through @xmath23 . for calculations of the flux one should integrate @xmath22 over the directions of the vector @xmath24 . this determines the total number of photons of a given energy which reach the point @xmath5 at the moment @xmath6 , [ ec4 ] g(k , r , t)= g(k , r , t ) d_n . for integration over @xmath25 , it is convenient to write eq . ( [ ec3 ] ) in the form [ ec5 ] g(,,t)=^_0 ds d^3r_0^ q(k_r , r_0^,t - s)(r - r_0 ^ -n s ) , where @xmath26 is the three - dimensional @xmath27-function , and @xmath28 . after a simple integration of eq . ( [ ec5 ] ) over @xmath29 , we apparently would return to eq . ( [ ec3 ] ) . in the integrand of eq . ( [ ec5 ] ) , the vector @xmath24 enters only in the argument of the @xmath27-function , therefore for determination of @xmath30 , one should calculate the integral [ ec6 ] x ( r - r_0 ^ -n s)d_n . using the equation ( r - r_1^)=(r - r_1^ ) ( -_1^)(-_1^ ) , where @xmath31 ( @xmath32 ) are the spherical coordinates of the point @xmath5 ( @xmath33 ) , we find [ ec6 ] x=. then the function @xmath30 becomes @xmath34 [ ec7 ] ( |-_0|-s ) , and the integration over @xmath4 results in [ ec8 ] g(k,,t)= d^3r_0 . this equation has a simple physical meaning and can be also obtained from heuristic considerations . from the definition of the function @xmath23 , it follows that [ ec9 ] ^2dis the number of photons in the unit volume at the point @xmath5 and in the energy interval @xmath35 , emitted by a point - like source located at the point @xmath36 . note that due to the delay , the number of photons at the instant @xmath6 is determined by the source function at the moment @xmath37 . in the case of an extended source , for calculation of the number of photons one should use the sum of eq . ( [ ec9 ] ) type expressions . the result will be the quantity @xmath38 . replacing the sum by the integral , we would arrive at eq . ( [ ec8 ] ) . ( [ ec8 ] ) is correct for any @xmath5 and @xmath6 . in order to calculate the photon distribution at a distance that significantly exceeds the characteristic size of the source @xmath39 , let s define the origin of coordinates somewhere inside the source . then for @xmath40 , one can set @xmath41 and @xmath42 . neglecting in the denominator @xmath43 compared to @xmath44 , we obtain [ gtilde ] g(k,,t)=1r^2 d^3 r_0 q(k , r_0,t - r+(n_rr_0 ) ) , where @xmath45 . this means that all photons detected by a distant observer travel essentially in the direction of @xmath46 . therefore below we will not distinguish between the vectors @xmath47 and @xmath41 , i.e. @xmath48 . on the other hand , we should note that in general one can not neglect the last term in the argument @xmath49 . the possibility of such an approximation is determined not by the relation of @xmath44 to @xmath50 , but by the speed of variation of the function @xmath23 during the time @xmath50 . let us consider the collision of ultrarelativistic electrons and soft photons . it is convenient to describe the ensemble of electrons by the distribution function in the phase space @xmath51 . by definition , the quantity @xmath52 is the number of electrons located at the moment of time @xmath6 in the phase volume element @xmath53 . let s define @xmath54 as the number of photons of energy @xmath55 within the interval @xmath56 in the unit volume . we assume that the distribution of these target photons is homogeneous , isotropic and stationary . then the high energy photons appearing due to ics are described by the source function [ ex_com1 ] q(k , r , t)= w(p,_ph , k)f_e(p , r , t ) n_ph(_ph)d^3pd_ph , where @xmath57 is the probability of scattering averaged over the directions of the soft photons . according to eq . ( [ gtilde ] ) , the corresponding expression for @xmath30 is given by @xmath58 [ com2 ] f_e(p , r_0,t - r+nr_0 ) n_ph(_ph ) . all quantities in eq . ( [ com2 ] ) are relevant to a reference system @xmath59 ( the observer system ) . let s assume that the accelerated electrons belong to a blob which moves with a relativistic speed @xmath60 and lorentz factor @xmath61 . in order to determine the impact of the blob s bulk motion on the energy distribution of secondary photons , we introduce a co - moving coordinate system @xmath62 , and define @xmath63 as the distribution function of electrons in the blob , i.e. in the @xmath62 system . below all quantities in the @xmath62 system will be indicated by the prime symbol . note that the distribution function in the phase space is a relativistic invariant ( @xcite , @xcite ) , i.e. [ com3 ] f(p , r , t)=f(p,r,t ) , where the `` primed '' and `` unprimed '' variables are connected via lorentz transformations . if we assume ( without a loss of generality ) that the source is moving along the @xmath64-axis , then eq . ( [ com3 ] ) can be written in the form [ com4 ] f(p , x , y , z , t)=f(p,x , y,(z - vt),(t - vz ) ) , where [ com5 ] p_x = p_x^,p_y = p_y^p_z=(p_z^-ve_e ) . in eq . ( [ com2 ] ) the function @xmath65 is given at an instant delayed in time . replacing in eq . ( [ com4 ] ) @xmath6 by @xmath66 , one finds @xmath67 @xmath68 @xmath69,\ ] ] where @xmath70 are the components of the unit vector @xmath24 . note that for simplicity we have set the retarded time as @xmath71 . let s introduce the new variables of integration in a way that the space arguments of the function @xmath72 can be written in the form @xmath73 . apparently , for this we should set & x_0=x_0,y_0=y_0 , & + & z_0=(z_0 ^ -v(+n_xx_0+n_yy_0+n_zz_0)).&[com7 ] as it follows from eq . ( [ com7 ] ) , @xmath74 ; therefore the volume element is transformed according to [ com8 ] d^3r_0==d^3 r_0 , where [ com9 ] = 1(1-v n_z)=1(1-nv ) is the doppler factor . note that the transformation given by eq . ( [ com8 ] ) differs from the standard lorentz transformation . at @xmath75 , we have an increase ( but not a contraction as in the case of lorentz transformation ) of the volume by a factor of @xmath1 . the time - argument in the right part of eq . ( [ com6 ] ) with respect to the new variables @xmath73 becomes [ com10 ] ( + n_x x_0+n_y y_0+(n_z - v)z_0 ) . let s introduce a unit vector along @xmath76 . using eq.([com5 ] ) , we obtain the following well - known expressions for the aberration of light : . [ com11 ] n_x = n_x^,n_y = n_y^,n_z=(n_z^-v ) . then the time - argument in eq . ( [ com6 ] ) can be written as + ( nr_0 ) . as a result , we find that the function @xmath30 in the system @xmath59 is expressed via @xmath72 in the system @xmath62 as @xmath77 [ com12 ] f_e(p,r_0,+n_rr_0 ) n_ph(_ph ) . remarkably , the function @xmath72 is not constrained by any condition ; the distribution of electrons in the comoving frame can be _ non - stationary _ , _ non - homogeneous _ , and _ anisotropic_. we should note that , in the case of a homogeneous , isotropic and stationary target photon field , the homogeneity of electrons becomes irrelevant , i.e. eq . ( [ com12 ] ) does not depend on the spatial distribution of electrons . ( [ com12 ] ) can be significantly simplified after the following approximations . let s assume that the distribution function of electrons in the comoving frame @xmath62 is stationary and isotropic , i.e. @xmath78 , where @xmath79 is the electron energy in the @xmath62 frame . the energy @xmath80 in the @xmath59 frame and @xmath79 are related as @xmath81 where @xmath82 is the electron speed . furthermore , we propose that the lorentz factor of electrons significantly exceeds the lorentz factor of the bulk motion , @xmath83 , and take into account that the up - scattered photon moves practically in the direction of the electron ( the accuracy of this approximation is of the order of @xmath84 ) . therefore , @xmath85 and , consequently , @xmath86 . integration of @xmath87 over all directions gives @xmath88 [ com14 ] = , where = , q=. for the given values of @xmath89 and @xmath80 , the maximum energy of the upscattered photon is _ ^=. usually eq . ( [ com14 ] ) is obtained by integration over the directions of the momentum of the upscattered photon ( @xcite and @xcite ) . however , the argument that depends on the angle enters in the integrand as @xmath90 , therefore there is no difference over the directions of which vectors , @xmath7 or @xmath91 , the integration is performed . the above formulae are applicable under the conditions _ ph m e_e . let s denote by @xmath92 the upscattered photon energy spectrum per electron : [ com15 ] w(e_e,)= w(e_e,_ph , ) n_ph(_ph ) d_ph . for the above simplifications , the observed energy flux ( @xmath93 ) becomes [ ecflux ] f_= n_e()w(e_e , ) de_e , where @xmath94 is the differential number of electrons in the comoving frame per energy and per solid angle . in the thomson limit ( @xmath95 ) , @xmath96 becomes a function of a single argument , @xmath97 . writing @xmath98 , and performing an integration over @xmath99 , we obtain [ ecflux1 ] f_= n_e(e_e ) ( ) de_e . this implies that if the ec flux of a source at rest is described by some function @xmath100 , the relativistic bulk motion of the source results in [ ecflux2 ] f_=^4s(/^2 ) . in a log - log plot the function @xmath101 is obtained from @xmath102 by moving the latter up by a factor of @xmath103 , and shifting it to the right by a factor of @xmath104 . the total intensity is enhanced by a factor of @xmath105 : [ ecflux3 ] _ 0^f_d=^6_0^s()d . integrating over all angles , we obtain the luminosity detected by the observer ( the apparent luminosity ) : [ ecflux4 ] l=15(16 ^ 4 - 12 ^ 2 + 1)l_0 , where @xmath106 is the luminosity of the source at rest ( the intrinsic luminosity ) . the above results are obtained for an arbitrary distribution of relativistic elections . for electrons with a power - law distribution , @xmath107 , in the thomson limit we arrive at the well known result for the beaming pattern , @xmath108 @xcite . in this section we study the impact of a possible anisotropy of the electron distribution inside the moving source on the angular and energy distributions of ec radiation . non - isotropic distribution of electrons might be caused by different reasons . in particular , anisotropic particle distributions are expected within different acceleration scenarios , including the particle acceleration by relativistic shocks ( see e.g. @xcite ) , by the converter mechanism @xcite or due to magnetic reconnection @xcite . we will use the general eq . ( [ com12 ] ) assuming that the source function ( distribution of electrons ) in the comoving system is not isotropic , but stationary . let s introduce anisotropy ( in @xmath62 ) in the form [ anis0 ] f(p,r)=(n_e)f(e_e , r ) , where @xmath109 . the calculations for anisotropy in a general form are quite complex , therefore , for the purpose of demonstration , we will use an empirical approach , namely , adopt the specific form of elongated ellipsoid of revolution described by the function [ anis1 ] ( n_e)= . here @xmath110 denotes a constant unit vector , which could have an arbitrary direction , and the parameter @xmath111 is confined within the interval @xmath112 . the function @xmath113 is normalized by the condition [ anis2 ] ( n_e)=1 . the angular distribution given by eq . ( [ anis1 ] ) is azimuthally symmetric relative to @xmath110 . if the axis @xmath64 is directed along @xmath110 , the function @xmath114 can be considered as an equation of the surface of the ellipsoid of revolution written in spherical coordinates . the semi - major axis of the ellipsoid is directed along @xmath110 and is equal to @xmath115 ; the another two semi - axes are @xmath116 . it is convenient to introduce the asymmetry parameter @xmath117 , [ anis3 ] = -1=1 - 1 . in the limit of @xmath118 , the ellipsoid degenerates into a sphere , i.e. the angular distribution becomes isotropic ( @xmath119 ) . at @xmath120 , the angular distribution is strongly extended in the directions of @xmath110 and @xmath121 . we allow @xmath117 to be function of @xmath79 , i.e. the anisotropy can be energy - dependent . in the case of anisotropic angular distribution , the integration of eq . ( [ com14 ] ) over directions of the electron momentum results in [ anis4 ] w_anis=^2(n_e ) w(p,_ph , k)d_e . since the function @xmath87 is different from zero when the directions of vectors @xmath91 and @xmath7 practically coincide , the argument of the function @xmath113 in eq . ( [ anis4 ] ) can be replaced by @xmath122 , and @xmath123 can be taken out under the sign of integral . this gives [ anis5 ] w_anis ( n ) w(e_e,_ph , ) , where @xmath122 in the argument of @xmath113 should be expressed via @xmath24 according to relations in eq . ( [ com11 ] ) . therefore , for the anisotropic angular distribution of electrons the observed flux is [ ecflux2 ] f_=(n ) n_e()w(e_e , ) de_e , it is interesting to compare this equation with eq . ( [ ecflux ] ) . in the thomson regime of scattering , eq . ( [ ecflux2 ] ) is simplified , [ ecflux3 ] f_=(n ) n_e(e_e ) ( ) de_e . in the case of an energy - independent anisotropy ( @xmath124 ) , for the total ( integrated over energy ) intensity of radiation in the given direction , we have [ d6 ] f_d(n)^6 . for illustration of the results , we fix the orientations of the axes in the following way . as before , the axis @xmath64 is directed along the velocity @xmath125 ; the axis @xmath126 we choose from the condition that the vector @xmath110 becomes parallel to the @xmath127 plane . then the components of the vector @xmath110 can be written in the following form s=(0,_*, _ * ) . here the angle @xmath128 determines the direction of the axis of symmetry of the angular distribution of electrons in the comoving frame . to simplify the analysis , we assume that the radiation is detected in the plane @xmath127 , and introduce polar coordinates in this plane . then @xmath129 , where @xmath130 varies in the interval from @xmath131 to @xmath132 ( @xmath133 for the points on the axis @xmath64 ) . expressing in this equation @xmath130 through the viewing angle @xmath134 in the system @xmath59 , we get ( sn)= ( _*+(-v) _ * ) . this expression is to be substituted into eq . ( [ anis1 ] ) . + in fig [ angle_dist ] we show the dependence of the intensity of the total ec radiation ( the flux integrated over the photon energies ) on @xmath134 - the angle between the line of sight and the direction of the jet . the calculations are performed for different combinations of the bulk motion lorentz factor @xmath135 and the asymmetry parameter @xmath117 . the curves shown in these figures correspond to five different directions of the axis of symmetry of the electrons angular distribution described by the angle @xmath128 . for comparison we show also the intensity for isotropically distributed electrons . one can see that except for @xmath136 and @xmath137 , the maximum of the observed radiation appears not at @xmath138 , as it happens in the isotropic case . instead , it is shifted to larger angles as the level of anisotropy increases . for the chosen parameter , the shift can be as large as several degree . this implies that in the case of anisotropic distributions of electrons in the source , we should be able to see misaligned jets as it has been indicated by @xcite . in a more general approach , the @xmath117-parameter can depend on the electron energy . this could be realized , for example , in the case of diffusive shock acceleration of electrons when the low - energy particles can be effectively isotropised in the downstream region due to pitch - angle scattering , whereas higher - energy particles radiate away their energy before being fully isotropised ( see , e.g. @xcite ) . to demonstrate this effect , let s assume a simple energy - dependence of the asymmetry parameter , @xmath139 , and consider a distribution of electrons in the standard `` power - law with an exponential cutoff '' form : [ anis6 ] n_e(,n ) = a ( n)^-pe^-/_0 , where @xmath140 . here instead of the electron energy @xmath79 , we use its lorentz factor , @xmath141 . from the condition of normalization in eq . ( [ anis2 ] ) it follows that @xmath142 does not depend on the level of anisotropy . in fig . [ ec_anis ] we show the spectral energy distribution of ec radiation assuming that a relativistic jet propagates through the 2.7 k microwave background radiation . this might be relevant to the inverse compton x - ray emission of extended jets of agn which could be ( still ) relativistic on kpc scales ( see , e.g. @xcite ) . the radiation spectra are calculated for the viewing angle @xmath138 , thus @xmath143 , and for the following combination of model parameters : @xmath144 , @xmath145 , @xmath146 and @xmath147 . since @xmath148 , the compton scattering proceeds in the thomson limit one can see from fig . [ ec_anis ] that if the photon energy is expressed in units of @xmath149 , the shape of the energy spectrum of ec radiation does not depend on the jet s doppler factor . on the other hand , it depends on the angle @xmath128 . the apparent reason is the dependence of the electron energy distribution on @xmath128 . for example , for @xmath138 and @xmath150 , the spectrum given by eq . ( [ anis6 ] ) for @xmath136 can be written in the form [ anis7 ] n_e(,n ) = ^-p+p_1e^-/_0 , for @xmath137 , eq . ( [ anis6 ] ) differs from the spectrum corresponding to the isotropic distribution of electrons , by a constant factor of @xmath151 . in various astrophysical objects , like microquasars , agn , and grbs , the observed fluxes of radiation emerge from relativistically moving jets . the doppler boosting caused by this motion can significantly enhance ( by orders of magnitude ! ) the emitted absolute flux and shift the spectrum towards higher energies . therefore , the beaming pattern of radiation is a key issue for proper understanding of acceleration and radiation processes in these objects . in this paper we derived the energy distribution of the ec radiation by solving the photon transfer equation for an optically thin source in a rather general case . it is described by eq . ( [ com12 ] ) which allows non - stationary and non - isotropic distribution of electrons in the frame of a relativistically moving source . ( [ com12 ] ) does not specify the energy distribution of electrons either , but requires isotropic , homogeneous and non - variable fields of seed photons for ics . the latter condition makes the solution independent of the spatial distribution of electrons . for power - law energy distribution of isotropically distributed electrons ( [ com12 ] ) is reduced to previously derived results @xcite . anisotropic distribution of electrons in a relativistically moving source can be realized in some acceleration scenarios , therefore it is of a special practical interest . the formalism developed in this paper has been used to study the impact of the electron anisotropy on the ec emission . the calculations show that the anisotropy of emitting particles can significantly modify the beaming pattern . most notably , the emission peak can be significantly shifted relative to the line of sight . this implies that , thanks to the anisotropic distribution of electrons in the source , modestly mis - aligned jets may become detectable . and vice versa , while the energy of strongly anisotropic distribution of electrons in a source at rest can be radiated away from the observer , the relativistic motion of the source would make the radiation detectable , even in the case of most unfavorable anisotropy of electrons . generally , the electron anisotropy is expected to be energy - dependent . in this case the anisotropy could result in harder spectra of ec emission compared to the isotropic distribution of electrons . the effects related to the anisotropic distribution of electrons in general , and in the context of the ec scenario , in particular , are quite strong . they can not be ignored when interpreting the high energy emission from highly relativistic jets in agn and grbs .

the beaming pattern of radiation emitted by a relativistically moving source like jets in microquasars , agn and grbs , is a key issue for understanding of acceleration and radiation processes in these objects . in this paper we introduce a formalism based on a solution of the photon transfer equation to study the beaming patterns for emission produced by electrons accelerated in the jet and upscattering photons of low - energy radiation fields of external origin ( the so - called external compton scenario ) . the formalism allows us to treat non - stationary , non - homogeneous and anisotropic distributions of electrons , but assuming homogeneous / isotropic and non - variable target photon fields . we demonstrate the non - negligible impact of the anisotropy in the electron distribution on angular and spectral characteristics of the ec radiation .

fractional differential equations ( fdes ) have become more and more popular in applied science and engineering field recently . the history and mathematical background of fractional differential operators are given in @xcite with definitions and applications of fractional calculus . this kind of equations has been used increasingly in many fields , for example , in nature @xcite fractional operators applied in fractal stream chemistry and its implications for contaminant transport in catchments , in @xcite the fractional calculus motivated into bioengineering , and its application as a model for physical phenomena exhibiting anomalous diffusion , l@xmath1vy motion , turbulence @xcite , etc . let us briefly review the development of numerical methods for the fractional convection - diffusion equations . several authors have proposed a variety of high - order finite difference schemes for solving time - fractional convection - diffusion equations , for example @xcite , and solving space - fractional convection - diffusion equations @xcite . in @xcite , w. mclean and k. mustapha have used the piecewise - constant and piecewise - linear discontinuous galerkin ( dg ) methods to solve the time - fractional diffusion and wave equations , respectively . but these methods require more computational costs . in order to tackle those problems , in @xcite w. mclean has proposed an efficient scheme called fast summation by interval clustering to reduce the implementation memory ; more recent works on this issue can been in @xcite . furthermore , in @xcite deng and hesthaven have developed dg methods for fractional spatial derivatives and given a fundamental frame to combine the dg methods with fractional operators . in @xcite xu and hesthaven have applied the dg methods to the fractional convection - diffusion equations in one dimension . in the two dimensional case , ji and tang @xcite have applied the dg methods to recast the fractional diffusion equations in rectangular meshes with the numerically optimal convergence order @xmath2 . however , there are no theoretical results . so far very few literatures deal with the fractional problems in triangular meshes , besides @xcite . this motives us to consider a successful dg method for solving the fractional problems in triangular meshes . here , we consider the time - dependent space - fractional convection - diffusion problem @xmath3 in the domain @xmath4 and @xmath5 $ ] with the superdiffusion operators which are defined by the left riemann - liouville fractional derivatives @xmath6 and @xmath7 , @xmath8 . the function @xmath9 is a source term ; the convection coefficient @xmath10 is supposed to satisfy @xmath11 , and the initial function @xmath12 . in this work , we shall design a stable and accurate dg method for ( [ eq1 ] ) . the stability and convergence are proved in multi - dimensional case . this development is built on the extension of the previous dg works found in @xcite , where a qualitative study of the high - order local dg methods was discussed and some theoretical results were offered in one space dimension . in order to perform the error analysis , the authors defined some projection operators to prove the convergence results . unfortunately , the defined projection operators can not be easily extended to two dimensional case ( see @xcite ) . hence , to avoid this difficulty , a different dg method is designed in this paper by carefully choosing the numerical fluxes and adding penalty terms . the presented hybridized discontinuous galerkin ( hdg ) method has the following attractive properties : 1 ) the hdg method can be used for other fractional problems , for example , fractional diffusion equations ; 2 ) it has excellent provable stability , i.e. , the stability can be proved in any space dimensions ; 3 ) theoretically , the error analysis can be more easily performed with the general analytical methods in any space dimensions . the outline of this paper is as follows . in section 2 , we introduce some basic definitions , notations and review a few lemmas which are useful for the following analysis . in section 3 , we present the computational schemes and give some discussions . in section 4 , we perform the stability and convergence analysis for the 2d space - fractional convection - diffusion equations . in section 5 , we make the numerical experiments and show some simulation results to verify the theoretical results and illustrate the performance of the proposed schemes . we conclude the paper with some remarks in the last section . in the following we give some definitions of fractional integrals , derivatives , and their properties . [ definition2.1 ] for any @xmath13 , the left and right riemann - liouville fractional integrals of function @xmath14 defined on @xmath15 are defined by @xmath16 and @xmath17 [ definition2.2 ] for any @xmath18 , the left and right riemann - liouville fractional derivatives of function @xmath19 defined on @xmath15 are defined by @xmath20 and @xmath21 [ definition2.5 ] for any @xmath18 , caputo s left and right fractional derivatives of function @xmath14 on @xmath15 are defined by @xmath22 and @xmath23 [ lemma2.3 ] for any @xmath13 , the left and right riemann - liouville fractional integral operators are adjoints for any functions @xmath24 , i.e. , @xmath25 [ lemma2.6 ] suppose that @xmath14 is a function defined on @xmath15 , @xmath26 when @xmath27 or @xmath28 , @xmath29 . there are @xmath30 or @xmath31 note that , from definition [ definition2.2 ] , definition [ definition2.5 ] and lemma [ lemma2.6 ] if the solution @xmath19 of ( [ eq1 ] ) satisfies @xmath32 when @xmath27 or @xmath33 , then for any @xmath8 , the left fractional riemann - liouville derivatives of function @xmath34 on @xmath4 can be rewritten as ( see @xcite ) : @xmath35 @xmath36 for the convenience , we use the notation @xmath37 where @xmath38 and @xmath39 . [ definition2.7 ] for @xmath40 , extend @xmath14 outside of @xmath41 by zero . define the norms @xmath42 @xmath43 let the two spaces @xmath44 and @xmath45 denote the closures of @xmath46 with respect to @xmath47 and @xmath48 , respectively . [ lemma2.8 ] for @xmath13 , assume that @xmath14 is a real function . then @xmath49 generally , we consider the case in which the problem is in a bounded domain instead of @xmath50 . thus we restrict the definitions to @xmath51 . [ definition2.9 ] define the spaces @xmath52 and @xmath53 as the closures of @xmath54 under their respective norms . [ theorem2.10 ] if @xmath55 , then @xmath56 and @xmath57 are embedded into @xmath58 and @xmath59 , respectively . furthermore , @xmath60 is embedded into both of them . [ definition2.11 ] by lemma [ lemma2.3 ] , lemma [ lemma2.8 ] , definition [ definition2.7 ] and definition [ definition2.9 ] , we obtain @xmath61 let the spaces @xmath62 and @xmath63 denote the closures of @xmath64 and @xmath65 under their respective norms , and @xmath66 , @xmath39 . we first review some notations , and then focus on deriving the fully discrete numerical scheme of the 2d space - fractional convection - diffusion equation . for the mathematical setting of the dg methods , we describe some spaces and notations . the domain @xmath67 is subdivided into elements @xmath68 . here @xmath68 is a triangle in 2d . we assume that the intersection of two elements is either empty , or an edge ( 2d ) . the mesh is called regular if @xmath69 where @xmath70 is the subdivision of @xmath67 , @xmath71 a constant , @xmath72 the diameter of the element @xmath68 , and @xmath73 the diameter of the inscribed circle in element @xmath68 . throughout this work @xmath74 . we introduce the broken sobolev space for any real numbers @xmath75 by @xmath76 equipped with the broken sobolev norm : @xmath77 the set of edges of the subdivision @xmath70 is denoted by @xmath78 . let @xmath79 denote the set of interior edges , and @xmath80 denote the set of edges on @xmath81 . with each edge @xmath82 , the unit normal vector is @xmath83 . if @xmath82 is on the boundary @xmath81 , then @xmath83 is taken to be the unit outward vector normal to @xmath81 @xcite . if @xmath84 belongs to @xmath85 , then the trace of @xmath84 along any side of one element @xmath68 is well defined . if two elements @xmath86 and @xmath87 are neighbours and share one common side @xmath82 , then there are two traces of @xmath84 belonging to @xmath82 . we assume that the normal vector @xmath83 is oriented from @xmath86 to @xmath87 . then the average and jump are defined , respectively , by @xmath88 if @xmath82 is on @xmath81 , we have @xmath89 for designing the dg method of fractional derivative , we rewrite ( [ eq1 ] ) as a low order system ( see @xcite ) . firstly , we introduce two auxiliary variables @xmath90 and set @xmath91 as ref . @xcite , let @xmath92 , where @xmath93 . hence , the characteristic direction associated with @xmath94 is denoted by @xmath95 . then , from ( [ equation2.4 ] ) and ( [ equation2.5 ] ) , eq . ( [ eq1 ] ) can be rewritten as a mixed form @xcite : @xmath96 where @xmath97 is defined in ( [ eq2.6 ] ) . for an arbitrary subset @xmath98 , we multiply the first , second , and the third equation of ( [ eqloworder ] ) by the smooth test functions @xmath99 , and @xmath100 , respectively . in order to obtain a symmetric weak variational formulation , we only integrate the first equation of ( [ eqloworder ] ) by parts , and obtain @xmath101 where @xmath102 is the outward unit normal to @xmath103 . note that the above equations are well defined for the functions @xmath104 and @xmath105 in @xmath106 , where @xmath107 next we will approximate the exact solution @xmath104 with the functions @xmath108 in the finite element spaces @xmath109 , where @xmath110 where the finite element space @xmath111 denotes the set of polynomials of degree less than or equal to @xmath112 . thus , the approximate solution @xmath108 satisfies the weak formulation , for all @xmath113 , @xmath114 where the numerical fluxes are well chosen as @xmath115 in order to ensure the stability of the scheme and its accuracy . it is well known that the fluxes @xmath116 are consistent . inspired by the penalty galerkin methods @xcite and noting the fact that @xmath117 and @xmath118 , a symmetric and stable dg scheme is derived as follows . substituting the flux @xmath116 into ( [ eq3.3 ] ) , summing over all the elements , and adding the penalty terms , we observe that for @xmath119 , the semi - discrete variational formulation is given by @xmath120 for any @xmath121 , the exact solution of ( [ eq1 ] ) is expected to be continuously differentiable with respect to the variables @xmath122 and @xmath123 , which keeps the consistency of the scheme . the term @xmath124 vanishes since the exact solution @xmath19 satisfies @xmath117 . note that @xmath125 penalizes the jump of the function @xmath19 , whereas @xmath126 penalizes the jump of the function @xmath127 . here @xmath128 and @xmath129 are the positive constants to be chosen . unfortunately the third equation of ( [ eq3.4 ] ) makes the dg method lose its locality , since @xmath130 is a function of @xmath131 and @xmath132 , @xmath130 can not be eliminated from the third equation . so we have to simultaneously solve the three unknowns @xmath133 . although the extra unknowns can not be eliminated in the hdg methods , our choice of fluxes makes the error analysis available . above and throughout , the following notations are used , @xmath134 after performing the hdg approximation , we discretize the time derivative with the characteristic method . for the given positive integer @xmath135 , let @xmath136 be a partition of @xmath137 into subintervals @xmath138 $ ] with uniform mesh and the interval length @xmath139 . the characteristic tracing back along the field @xmath10 of a point @xmath140 at time @xmath141 to @xmath142 is approximated by @xcite @xmath143 therefore , the approximation for the hyperbolic part of ( [ eq1 ] ) at time @xmath141 can be approximated as @xmath144 where @xmath145 , @xmath146 , and @xmath147 . [ see @xcite][remark 2 ] assume that the solution @xmath19 of ( [ eq1 ] ) is sufficiently regular . under the assumption of the function @xmath10 , we have @xmath148 thus , the fully discrete scheme corresponding to the variational formulation ( [ eq3.4 ] ) is to find @xmath149 , for any @xmath150 , such that @xmath151 where @xmath152 . define the bilinear forms by @xmath153 and the linear form @xmath154 we can rewrite ( [ eq3.5 ] ) as a compact formulation : find @xmath149 at time @xmath155 , such that @xmath156 this section focuses on providing the proof of the unconditional stability and the error estimates of the schemes . in the following , @xmath71 indicates a generic constant independent of @xmath157 and @xmath158 , which takes different values in different occurrences . [ @xcite][lemma4.1 ] if @xmath11 , for any function @xmath159 and each @xmath160 , there is @xmath161 where @xmath162 . [ theorem4.2 ] if @xmath11 , the hdg scheme ( [ eq3.5 ] ) is stable , i.e. , for any integer @xmath163 , there is @xmath164 where @xmath165 , and the semi - norm @xmath166 is defined as @xmath167 let @xmath168 in the equations of ( [ eq3.6 ] ) , respectively . by the symmetry of the bilinear formulas , adding the above equations , we obtain @xmath169 following from @xmath170 the young inequality , the definition of @xmath171 and @xmath166 , and lemma [ lemma4.1 ] , we have @xmath172 summing from n=1,2, ... ,n , we get @xmath173 using the discrete grnwall inequality , with @xmath174 , there is @xmath175 in this subsection we state and discuss the error bounds for the hdg scheme . the main steps of our error analysis follow the classical methods in finite element analysis , i.e. , the so - called galerkin orthogonality property . as usual , we denote the errors @xmath176 by @xmath177 where @xmath178 and @xmath179 are the @xmath180-projection and @xmath181-projection operators from @xmath182 and @xmath183 onto the finite element spaces @xmath184 and @xmath185 , respectively . from ( [ eq3.6 ] ) , we obtain the compact form @xmath186 where @xmath187 assume that the solution @xmath19 of problem ( [ eq1 ] ) is sufficiently regular . then @xmath188 by the consistency of the numerical fluxes , the exact solution @xmath104 satisfies ( [ eq3.4 ] ) . taking @xmath189 and subtracting ( [ eq3.5 ] ) from ( [ eq3.4 ] ) yield @xmath190 and @xmath191 by the galerkin orthogonality , there is @xmath192 substituting the equalities ( [ erro2 ] ) and ( [ erro3 ] ) into ( [ erro1 ] ) leads to the desired result . next we review two lemmas for our analysis . the first one is the standard approximation result for the @xmath180-projection operator @xmath178 from @xmath193 onto @xmath194 satisfying @xmath195 for any @xmath196 . the second one is the standard trace inequality . [ @xcite][lemma4.3 ] let @xmath197 . @xmath178 is the @xmath180-projection operator from @xmath193 onto @xmath198 such that @xmath195 for any @xmath196 . then , for @xmath199 @xmath200 [ lemma4.4 ] there exists a generic constant @xmath71 being independent of @xmath72 , for any @xmath201 , such that @xmath202 now we are ready to prove our main results . in this subsection , we estimate the first left - side term of ( [ erro ] ) . [ lemma4.5 ] if @xmath11 , for any function @xmath203 and each @xmath160 , @xmath204 where @xmath205 . the following result is a straightforward consequence of the estimate of the first left - side term of ( [ erro ] ) . [ theorem4.6 ] assume that the solution @xmath19 of problem ( [ eq1 ] ) is sufficiently smooth and @xmath206 satisfies ( [ eq3.5 ] ) . if @xmath11 , we have @xmath207 from ( [ erro ] ) , it can be noted that @xmath208 using lemma [ lemma4.1 ] , we obtain @xmath209 where @xmath210 . also by the taylor expansion and the hlder inequality , there are @xmath211 and @xmath212 where @xmath213 and @xmath214 follow from cauchy - schwarz s inequality , young s inequality and lemma [ lemma4.5 ] . substituting @xmath215 into ( [ eq4.16 ] ) , the desired result is reached . in this subsection , we use the general analytic methods to get the bound of the right side term of ( [ erro ] ) . [ theorem4.7 ] let @xmath19 be sufficiently smooth solution of ( [ eqloworder ] ) . @xmath216 are standard @xmath180-projection operators of @xmath217 , and @xmath218 solve ( [ eq3.5 ] ) . if @xmath11 , we have @xmath219 from the definition of @xmath220 , we have @xmath221 using hlder s , young s inequalities and lemma [ lemma4.3 ] , we obtain @xmath222 from lemma [ lemma2.3 ] , lemma [ lemma4.3 ] , definition [ definition2.7 ] , definition [ definition2.11 ] , and theorem [ theorem2.10 ] , it follows that @xmath223 where @xmath224 and @xmath225 are chosen as sufficiently small numbers such that @xmath226 and @xmath227 . integrating the first term of @xmath228 by parts , and using the orthogonal property of projection operator @xmath229 , we get @xmath230 with the same deduction of @xmath231 , there is @xmath232 by lemma [ lemma4.3 ] , we get @xmath233 note that @xmath234 and @xmath235 vanish because of the orthogonal property of the projection @xmath229 . substituting @xmath236 into ( [ eq4.18 ] ) , the desired result is obtained . assuming that the solution of ( [ eq1 ] ) is sufficiently regular , we have the following error estimates . [ theorem4.8 ] let @xmath217 be the exact solution of ( [ eqloworder ] ) , @xmath237 the numerical solution of the fully discrete hdg scheme ( [ eq3.5 ] ) . if @xmath11 , for any integer @xmath163 , there is @xmath238 where @xmath239 , @xmath240 , @xmath241 , @xmath224 and @xmath225 are chosen as above , @xmath242 is dependent of @xmath243 . substituting the results of theorem [ theorem4.6 ] and theorem [ theorem4.7 ] into ( [ erro ] ) , there is @xmath244 with @xmath245 , multiplying the above inequality by @xmath246 on both sides , summing over @xmath160 from @xmath247 to @xmath135 , and using the discrete grnwall inequality , there is @xmath248 by the triangle inequality , we obtain the desired result . in this section , we illustrate the numerical performance of the proposed schemes by the numerical simulations of two examples . in the first example , we take the vector function @xmath249 and verify the accuracy of the schemes with the exact smooth solution @xmath19 combining with the left fractional riemann - liouville derivatives with respect to @xmath122-variable and @xmath123-variable , respectively . when we compute the fractional integral part in triangular meshes ( see figures 1 - 2 ) , the gauss points and weights are used to deal with the terms relating with the fractional operators element - by - element ( see @xcite ) . since this part needs more time and memory spaces ( see @xcite ) , we only use the piecewise linear basis functions to simulate the solution in triangular meshes . tables 1 - 3 illustrate that the schemes have a good convergence order with piecewise linear basis function for different choices of the fluxes . in the second example , we take @xmath10 to be a vector function and perform some numerical experiments with some figures ( see figures 3 - 4 ) which justify that the schemes simulate the solution very well for 2d - fractional convection - diffusion problems . @xmath250 consider 2d space - fractional convection - diffusion problem ( [ eq1 ] ) in domain @xmath251 . the initial condition and the exact solution are specified as @xmath252 then the force term @xmath253 is determined accordingly from ( 1.1 ) . in this case , we present a few results to numerically validate the analysis . for the numerical simulations , in order to validate the stability and the accuracy of the presented hdg scheme , we choose the time - stepsize , @xmath254 , used to advance the discrete formulation from @xmath142 to @xmath255 . the experimental convergence rate is given by @xmath256 .the @xmath257-errors and convergence rates for @xmath19 and @xmath258 for example 5.1 . [ cols="^,^,^,^,^,^,^,^ " , ] in table 1 and table 2 we choose different observation time @xmath259 and @xmath260 to justify that the convergence rates at least have an order of @xmath261 for the solution @xmath19 in @xmath262-norms based on the piecewise linear basis function . in table 3 we take the same choice of @xmath263 as ref . @xcite and see that the convergence rates increase to @xmath264 ( see the explanations in ref . @xcite ) . comparing the numerical results with the work @xcite , we can see that the hdg method has smaller numerical errors for the first order polynomial approximation . * example 5.2 . * in this example , we investigate the approximation solution of problem ( [ eq1 ] ) . for convenience , we still choose the domain @xmath251 . the exact solution @xmath19 , initial value and the vector function @xmath10 are given by @xmath265 for the second example , in order to further support the theoretical convergence and justify the powerful hdg scheme , we take @xmath10 to be nonzero vector function and give some approximation solutions with the refining space - step @xmath157 to compare with the exact solutions and display the efficiency of the simulations . figure 3 displays the exact solution @xmath19 and the numerical solutions @xmath131 based on different space stepsizes @xmath266 at @xmath267 with @xmath268 , @xmath269 . figure 4 displays the exact solution @xmath19 and the numerical solutions @xmath131 based on different space stepsizes @xmath266 at @xmath270 with @xmath271 . it is clear that the exact solution of example 5.2 is nonnegative with four hills . in the simulations , the @xmath272-hdg solutions recover the exact solution perfectly with all four hills in coarse meshes . note that the numerical results display that the approximations are more and more accurate with the refining of the meshes . by carefully introducing the auxiliary variables , constructing the numerical fluxes , adding the penalty terms , and using the characteristic method to deal with the time derivative and convective term , we design the effective hdg schemes to solve 2d space - fractional convection - diffusion equations with triangular meshes . as we know , this work is the first time to deal two - dimensional space - fractional convection - diffusion equation with triangular mesh by the dg method . the stability and error bounds analysis are investigated . besides the general advantages of hdg method , the presented scheme is shown to have the following benefits : 1 ) it is symmetric , so easy to deal with the fractional operators ; 2 ) theoretically , the stability can be more easily proved ; 3 ) the penalty terms make the error analysis more convenient ; 4 ) numerically verified to have efficient approximations ; 5 ) the schemes are performed very well in triangular meshes ; 6 ) it is possible to use this scheme to solve nonlinear equations which is the future research task . this work was partially supported by the national basic research ( 973 ) program of china under grant 2011cb706903 , the national natural science foundation of china under grant 11271173 and 11471150 , and the capes and cnpq in brazil . douglas , j. , russell , jr . , t. f. : numerical method for convection - dominated diffusion problems based on combining the method of characteristics with finite element or finite difference procedures . siam j. numer . 19*(5 ) , 871 - 885 ( 1982 ) . ji , x. , tang , h. z. : high - order accurate runge - kutta ( local ) discontinuous galerkin methods for one- and two - dimensional fractional diffusion equations . 5*(3 ) , 333 - 358 ( 2012 ) . liu , f. , zhuang , p. , anh , v. , turner , i. , burrage , k. : stability and convergence of the difference methods for the space - time fractional advection - diffusion equation . comp . * 191 * , 12 - 20 ( 2007 ) .

a hybridized discontinuous galerkin method is proposed for solving 2d fractional convection - diffusion equations containing derivatives of fractional order in space on a finite domain . the riemann - liouville derivative is used for the spatial derivative . combining the characteristic method and the hybridized discontinuous galerkin method , the symmetric variational formulation is constructed . the stability of the presented scheme is proved . theoretically , the order of @xmath0 is established for the corresponding models and numerically the better convergence rates are detected by carefully choosing the numerical fluxes . extensive numerical experiments are performed to illustrate the performance of the proposed schemes . the first numerical example is to display the convergence orders , while the second one justifies the benefits of the schemes . both are tested with triangular meshes . + + * ams * 26a33 , 35r11 , 65m60 , 65m12 .

we would like to thank the whole ganil and , in particular , the lise staff and the daq group for their support during the experiment . this work was partly funded by the conseil rgional daquitaine .

the decay of the ground - state two - proton emitter @xmath0fe was studied with a time - projection chamber and the emission of two protons was unambiguously identified . the total decay energy and the half - life measured in this work agree with the results from previous experiments . the present result constitutes the first direct observation of the individual protons in the two - proton decay of a long - lived ground - state emitter . in parallel , we identified for the first time directly two - proton emission from @xmath1cr , a known @xmath2-delayed two - proton emitter . the technique developped in the present work opens the way to a detailed study of the mechanism of ground - state as well as @xmath2-delayed two - proton radioactivity . since the advent of machines to produce atomic nuclei far away from the valley of nuclear stability , @xmath2-delayed particle emission and direct one - proton emission were among the tools most efficiently used to elucidate the structure of the atomic nucleus close to the proton drip line . the study of these decay modes allowed one to investigate the nuclear mass surface , to determine half - lives , to establish the sequence of single - particle levels or to improve the modeling of astrophysical processes . for the most proton - rich nuclei with an even number of protons , goldanskii @xcite predicted the occurence of a new nuclear decay mode , which he termed two - proton ( 2p ) radioactivity . he suggested that , at the proton drip line and due to the pairing of protons , nuclei exist , which are bound with respect to one - proton emission , but unbound for two - proton emission . in the early 1960 s , it was not clear which nuclei would be good candidates for this new decay mode . it is interesting that the list of possible 2p emitters proposed by goldanskii @xcite in the 1960 s contained already @xmath0fe , for which ground - state two - proton radioactivity was indeed discovered @xcite only recently . this new type of radioactivity was observed in projectile - fragmentation experiments at the lise3 separator of ganil @xcite and at the frs of gsi @xcite . these experiments allowed for an unambiguous identification of this decay mode by measuring the decay energy , which nicely fits modern decay @xmath3 value predictions @xcite , and the half - life , which is in agreement with models linking decay energy and coulomb and centrifugal barrier penetration half - lives @xcite . in addition , both experiments could demonstrate that , with a very large likelihood , there was no @xmath2 particle emission . this conclusion was achieved by a direct search for positrons @xcite or for the 511 kev photons from positron annihilation @xcite as well as by the width of the 2p emission peak , which is too narrow for a pile - up event of protons and @xmath2 particles . the most convincing piece of evidence is certainly the observation of the daughter decay with a half - life in nice agreement with the half - life of @xmath1cr , the 2p daughter of @xmath0fe @xcite . all other possible daughters ( e.g. @xmath2-delayed proton emission daughter ) can be excluded based on the experimentally measured daughter half - life . after the observation of 2p radioactivity for @xmath0fe , this decay mode was also identified for @xmath4zn @xcite and evidence has been found for 2p radioactivity of @xmath5ni @xcite . however , in both experiments the decay was observed in a silicon detector , which allows only for a determination of the total decay energy , the half - life , the branching ratio , and the absence of @xmath2 radiation , as in the case of @xmath0fe . in none of these experiments , the two protons were explicitly identified . emission of two protons has been observed directly only from excited states ( see e.g. @xcite ) or from very short lived resonances @xcite . in the present work , we report for the first time on the direct observation of two protons emitted by @xmath0fe . this is the first case of a direct two - proton observation of a long - lived ground - state two - proton emitter . the technique developped in the present work , the purpose of which is to determine the complete kinematics for two protons in three dimensions , paves the way for a detailed study of ground - state and @xmath2-delayed two - proton emission . it will enable us to investigate the decay dynamics , i.e. whether the decay is an ordinary three - body decay or whether the two protons are correlated in energy and in angle , the sequence of single - particle levels beyond the drip line and more . however , this nuclear structure information can only be extracted by a comparison to theoretical models . therefore , a development in parallel of powerful nuclear structure and reaction models is essential . the @xmath0fe nuclei were produced at the sissi - alpha - lise3 facility of ganil . a primary @xmath6ni@xmath7 beam with an energy of 75 mev / nucleon and an average intensity of 3 @xmath8a was fragmented in a @xmath9ni target ( 200@xmath8 m ) in the sissi device . the fragments of interest were selected by a magnetic - rigidity , an energy - loss , and a velocity analysis by means of the alpha spectrometer and the lise3 separator including an energy degrader ( 500@xmath8 m of beryllium ) in the lise3 dispersive focal plane and transported to the lise3 final focal plane . time - of - flight , energy - loss , and residual energy measurements allowed for an event - by - event identification of individual fragments . details of the identification procedure can be found in reference @xcite . m and a second position - sensitive detector ( 150@xmath8 m ) with resistive readout ) and are then implanted in the active volume of a gas - filled chamber , where they decay by emission of charged particles . the charge cloud created by the energy loss of either the heavy ions or the decay products drifts in the electric field of the chamber towards a set of four gas electron multipliers ( gems , not shown ) , where the charges are multiplied . the electric potential finally directs the charges onto a two - dimensional detection plane , where two orthogonal sets of 768 strips ( pitch of 200@xmath8 m ) allow for a measurement of the charges deposited on each strip . every second strip is equipped with an asic readout yielding signal height and time stamp . the other half of the strips is connected in groups of 64 strips to standard preamplifiers and shapers . ] fragments transmitted to the lise3 final focal plane were finally implanted in the center of newly developed time - projection chamber ( tpc ) @xcite , which allows for a visualisation in three dimensions of implantation and decay events ( see figure [ fig : tpc ] ) . for this purpose , the charges created by the primary ionising particles ( protons or heavy ions ) in the gas ( argon ( 90% ) - methane ( 10% ) ) drift in an electric field towards a set of four gas electron multipliers ( gems ) @xcite , where the charges are multiplied , and are finally detected by two orthogonal sets of 768 strips with a pitch of 200@xmath8 m . the front - end electronics , based on asic technology , allows for the measurement of the charge deposited on each strip and of the arrival time of the charge on this strip with respect to a start triggered by the first strip that fires . the asics were readout by vme multiplexed adcs . this readout mode yielded a dead time per event of about 1.3 ms . the signals from different strips were gain - matched first by injecting a pulser signal into the last gem thus creating an image charge on each single strip . this gain matching was refined by means of primary and fragment beams traversing the whole chamber with sufficient energy so that the energy loss can be assumed constant over the chamber . the strips parallel to the beam direction were gain - matched by rotating the chamber by 90@xmath10 . the tpc was extensively tested off - line with @xmath11 sources and on - line with different beams , which consituted a proof of the working principle of the chamber . details of this new instrument will be published elsewhere @xcite . after a setting on @xmath12ni , a well known @xmath2-delayed proton emitter with decay energies around 1.2 mev ( see e.g. @xcite ) , the main setting of the sissi - alpha - lise3 facility was optimised for the production and selection of @xmath0fe . in addition to @xmath0fe , a large number of @xmath1cr nuclei were also produced and implanted in the tpc in the same setting . @xmath1cr is a @xmath2-delayed two - proton ( @xmath22p ) emitter , for which , however , the two protons were never directly observed @xcite . in the present paper , we will present only the results obtained from the energy signals of the strips , which allows for an unambiguous identification of the two protons emitted by @xmath0fe . the analysis of the time signals , which will give rise to a three - dimensional reconstruction of the proton tracks , is much more complicated and is beyond the scope of the present paper . the complete analysis and the final results of the present experiment will be published in a subsequent paper . in order to accept a pair of implantation and decay events , the decay tracks in x and y have to start where the implantation tracks end . figure [ fig : fe45_implant ] shows an implantation event for @xmath0fe . the beam enters parallel to the x strips and stops at a certain y depth in the active volume of the tpc ( see figure [ fig : fe45_implant ] ) . there , the decay tracks of the two protons emitted have to start . fe . the ion enters parallel to the x strips ( left figure ) and stops at a certain depth in y direction ( right figure ) . this plot allows to determine the stopping position of the ion , where then the decay has to occur . lower part : the decay of @xmath0fe takes place by two - proton emission . the decay tracks start , within the position resolution of about 6 - 7 mm ( fwhm ) , at the stopping point of the implantation event described above . the arrows are indicative only of the tracks of the two protons . a horizontal arrow means that the proton propagates perpendicular to the strips , whereas a vertical arrow indicates that the proton track is parallel to the strips . the shaded zone corresponds to the error bars due to the gain matching procedure . ] in the present experiment , we observed a total of 10 @xmath0fe implantations , which could be correlated with decays . about 5 other decay events are lost to a large extent due to the data acquisition dead time mentioned above and the short half - life of @xmath0fe ( @xmath13 = 1.75@xmath14ms ) @xcite . in these cases , the treatment of the @xmath0fe implantation event prevents the registration of its decay . four @xmath0fe decay events are shown in the lower part of figure [ fig : fe45_implant ] and in figure [ fig : fe45_events ] . we will shortly explain the event in figure [ fig : fe45_implant ] in the following . on the lower left panel ( x strips ) , one sees the starting point of the two protons around channel 200 , in agreement with the maximum of the implantation event in figure [ fig : fe45_implant ] , and traces extending on the left and the right of this point , testimony of the two protons being emitted across a number of strips . the lower right panel indicates a starting point around strip 160 . one observes a large bump on which is superimposed a narrow one . the first indicates that one proton flies more or less perpendicular to the y strips , whereas the second proton has a trajectory paralell to these strips . therefore , this figure clearly evidences the presence of two charged particles in the decay of @xmath0fe . there is no other explanation possible for such a decay pattern . fe . we have chosen here events , where the two protons can be nicely identifed . for two of the ten events , the two protons can not be separated , as they fire the same strips . ] this is the first direct observation of the emission of two protons in the ground - state decay of a long - lived two - proton emitter . as mentioned above , all previous evidence , even if equally conclusive , was indirect . although the third coordinate z is not yet determined - it will follow from the drift - time analysis , one can already state at this stage that a variety of decay configurations is observed and not only one type , e.g. a back - to - back emission . after background subtraction , the charges deposited by decay events on the different strips can be integrated and are proportional to the energy of the decay event . charges on x and y strips were integrated separately . in addition , the different gems yield also the total decay energy . therefore , the charge signal from the gems is sent to a charge - sensitive pre - amplifier and a shaper . in figure [ fig : fe45_decay]a , we plot this charge signal from the last gem . a rough energy calibration was achieved by means of a triple @xmath11 source . the decay energy thus determined corresponds roughly to the expected decay @xmath3 value of 1.15 mev @xcite . the time difference between an implantation event and its subsequent decay event allows for a determination of the half - life of @xmath0fe . this spectrum is shown in figure [ fig : fe45_decay]b . a fit with an exponential , excluding the first 2 ms which are affected by the data acquisition dead time , yields a half - life value of ( [email protected] ) ms . this value is nice agreement with the average value from previous experiments of 1.75@xmath14 ms @xcite . fe events as determined with the last gem . the low - energy events ( in black ) show the @xmath0fe decay , whereas the high - energy events are from a triple-@xmath11 source . the energy of the @xmath0fe events correspond roughly to the expected decay energy of 1.15 mev @xcite . ( b ) half - life of @xmath0fe as determined from the time difference of the implantation events and the decay events ( black events ) . the half - life is ( [email protected] ) ms . the events in gray are the daughter - decay events which were not included in the fit . ] as mentioned above , the setting of lise3 on @xmath0fe allowed us also to study the decay of @xmath1cr , a known @xmath2-delayed two - proton emitter , for which the @xmath22p decay was only identified by means of the total decay energy @xcite . the present experiment with the tpc enabled us for the first time also to visualise the emission of two protons from this nucleus . although the active volume of the present version of the tpc is too small to stop all the protons in this decay due to their much higher energy ( up to about 2 mev for each proton ) , the two protons are nonetheless clearly observable for most of the decay events . one of these decay events together with its preceding implantation is shown in figure [ fig : cr43 ] . the two - proton tracks are visible in both directions . cr nucleus in the tpc . the nucleus is identified by means of energy - loss and time - of - flight measurements . lower part : decay of @xmath1cr as observed with the tpc . the traces of the two protons can be seen in the x and y direction . ] a detailed study of this decay and the search for a possible angular correlation between the two protons can be envisaged either by increasing the active volume of the tpc or by working at higher pressure . in this way , many other @xmath22p emitters could be identified and studied . in summary , we have performed the first experiment , which demonstrates unambiguously that @xmath0fe emits indeed two protons in most of its decays . this is the first time that ground - state two - proton emission of a long - lived ( longer than 10@xmath16s ) isotope is observed directly . thus , ground - state two - proton radioactivity is definitively established as a nuclear decay mode . a more detailed analysis will try to determine how the two protons share the energy and thus to elucidate the decay mechanism which governs two - proton radioactivity . the complete analysis of the energy and of the time signals should allow for a more detailed access to the decay kinematics . this will enable us to determine the proton - proton angle which then can be compared to model predictions for the decay pattern of two - proton radioactivity . however , to establish a detailed picture of the decay process , higher - statistics data ( 50 - 100 implantation and decay events ) are needed , which can be obtained in a future experiment .

let @xmath12 be a calabi - yau threefold , i.e. , a compact khler manifold with trivial canonical bundle and no holomorphic @xmath13-forms . as an example , take a generic anticanonical section @xmath6 of a smooth toric fano fourfold @xmath0 . following batyrev s seminal article ( @xcite ) , there exists a mirror partner of @xmath6 , which is given as follows . the toric fourfold @xmath0 is associated with a polytope in four - dimensional real space . the dual polytope yields another toric variety @xmath4 , which is not smooth in general . take a toric resolution @xmath3 associated to a maximal projective subdivision the fan of @xmath4 . a generic anticanonical section @xmath7 of @xmath3 is a mirror partner of a generic anticanonical section of @xmath0 . here by a mirror partner we mean only a hodge - theoretic mirror partner , i.e. , the relevant hodge numbers of @xmath6 and @xmath7 are exchanged , namely : @xmath14 let @xmath15 be a finite group of automorphisms of @xmath0 that acts freely on @xmath6 . it is easy to check that @xmath16 is a calabi - yau threefold with non - trivial fundamental group . it is natural to ask whether , for each pair @xmath9 , there exists a finite group @xmath17 such that @xmath18 is a mirror partner of @xmath8 - at least hodge - theoretically . in section ( [ sec : anti ] ) , we give a negative answer to the question above if the groups acts with the highest possible order . more precisely , let @xmath19 and consider pairs @xmath9 , where @xmath6 is a calabi - yau threefold in @xmath0 and @xmath15 is a finite group of automorphisms of @xmath0 that acts freely on @xmath6 with the maximum possible order . these pairs are first investigated in ( @xcite ) , where they are called _ maximal admissible pairs_. for each of them there exists a calabi - yau threefold @xmath20 with non - trivial fundamental group and relevant hodge numbers @xmath21 . a ( hodge - theoretic ) mirror partner of @xmath8 should have @xmath22 and height @xmath23 , where the height is defined as the sum of the two hodge numbers above . in this setting , @xmath0 is the toric fano fourfold associated with the hypercube . its dual is the @xmath24-cells ( or hyperoctahedron ) that yields a singular toric variety . if we take a maximal projective subdivision of the hypercube , we have a smooth toric variety , which is not fano because the anticanonical bundle is semi - ample and not ample ; however , this resolution is crepant . in other words , an anticanonical section @xmath7 of the resolution is a smooth calabi - yau , which is a mirror of @xmath6 . in order to find subgroups of @xmath25 , we recall the structure of the automorphism group of a toric variety ( cfr . this is given by the combinatorics of the fan , as well as by the dense torus in @xmath3 . the former yields a unique possible group in order to satisfy ( ) ; the latter helps us to describe all possible families of finite groups acting on @xmath7 . none of them acts without fixed points ; furthermore , the fixed locus is the union of some curves . therefore , a mirror - if it exists - can not be found in this way . of course , the quotients by the groups described above have finite quotient singularities , and there exists a mirror calabi - yau orbifold that satisfies ( ) . if we take an crepant resolution of the orbifold mentioned before , we get a smooth calabi - yau manifold with height @xmath26 and @xmath27 . this is a mirror partner ( at least hodge - theoretically ) of a quotient in ( @xcite ) , which is a calabi - yau manifold with cyclic fundamental group of order four : see remark ( [ osette ] ) . finally , in section ( [ sec : mir ] ) , we prove that it is indeed possible to find hodge - theoretic mirror pairs by taking quotients associated to admissible pairs in @xmath0 and in @xmath3 if the admissible pair in @xmath0 is not maximal . more precisely , we construct three pair @xmath28 of hodge - theoretic calabi - yau mirrors with @xmath29 associated to an admissible pair in @xmath0 and @xmath30 associated to an admissible pair in @xmath3 . * acknowledgments . * this work is part of the second named author s ph.d thesis . we are grateful to the anonymous referee of the thesis for suggestions and helpful remarks , especially on the chapter related to the material presented here . the first named author is partially supported by firb 2012 `` moduli spaces and applications '' and prin 2010 - 2011 `` geometry of algebraic varieties '' . both authors are partially supported by miur and gnsaga . let @xmath0 be @xmath1 . @xmath0 is a fano fourfold and it is the toric variety associated to the polytope in @xmath31-dimensional real space @xmath32 ^ 4 $ ] , the hypercube . the fan @xmath33 of @xmath0 is spanned by the faces of the dual polytope @xmath5 of @xmath2 , i.e. , the @xmath34cells that can be realized as the convex hull of the points @xmath35 , where @xmath36 form the canonical basis of @xmath37 . cells . ] as shown in ( @xcite ) and ( @xcite ) , there exists a smooth and @xmath38invariant anticanonical section @xmath6 with @xmath15 of order @xmath24 that acts freely on @xmath6 . the fano variety dual to @xmath0 is associated to the fan spanned by the faces of @xmath2 , i.e. , the hypercube . unfortunately , it is a strongly singular variety : indeed , its fan @xmath39 is neither smooth nor simplicial . for example , the cone whose primitive generators are @xmath40 is not simplicial and the same is true for all the cones of dimension @xmath41 and @xmath31 . in order to describe the mirror calabi - yau of @xmath6 , it is necessary to choose a maximal projective subdivision @xmath42 of @xmath39 . by definition , every maximal projective subdivision fulfills the following conditions : @xmath43 there are @xmath44 integral points in @xmath2 : @xmath24 vertices of the hypercube ( which correspond to the rays of @xmath33 ) , @xmath45 points that are in the middle of an edge , @xmath46 that correspond to centers of @xmath47faces ( squares ) , @xmath48 that are centers of one of the @xmath48 cubes of the hypercube ; finally , we have the center of the hypercube which is the origin . this implies that every maximal subdivision of @xmath39 will have @xmath49 rays . given a toric variety @xmath0 associated to a fan @xmath33 , we recall that its homogeneous coordinate ring @xmath50 is the ring @xmath51 $ ] with a particular grading on @xmath52 . if @xmath53 , we define for brevity @xmath54 . the grading on @xmath50 is given by the condition @xmath55 one of the most important thing about @xmath50 is that it only depends on the rays of the fan @xmath33 . different fans with the same rays give the same homogeneous coordinate ring . we recall that a root of @xmath0 is a pair @xmath56 such that @xmath57 is different from @xmath58 but they have the same degree . for each root it is possible to find a @xmath59parameter group of automorphisms of the corresponding toric varieties . by definition , another family of automorphisms of toric varieties are those coming from the torus @xmath12 . if @xmath0 is simplicial , there is only another elementary type of automorphisms , those obtained by the symmetries of the fan @xmath33 . we denote this group by @xmath60 , where @xmath61 is the lattice such that @xmath62 is the vector space where the fan lies . for simplicial varieties there is an exact sequence @xmath63 due to demazure ( smooth case ) and cox ( simplicial case ) . @xmath64 is a permutation group on the partition of @xmath65 given by the equivalence relation @xmath66 iff @xmath67 . assume that @xmath42 is a maximal projective resolution of @xmath39 and call @xmath3 the corresponding toric variety . then , @xmath25 is an algebraic group of dimension @xmath31 whose connected component that contains the identity is the torus @xmath12 , i.e. , @xmath3 has no roots . moreover , the exact sequence ( [ eq : demazure ] ) becomes @xmath68 it suffices to show that there are no roots . standard facts of toric geometry imply that the number of roots of a simplicial and gorenstein variety is given by the sum over the facet @xmath69 of @xmath70 of @xmath71 , i.e. , the number of integral points in @xmath69 that are not on its topological border . @xmath42 is a maximal projective subdivision and this implies that the polytope associated to the divisor @xmath72 is the polytope of @xmath4 . it is easy to see that the @xmath34cells . ] contain no integral points besides its vertices and the origin . this implies that there are no roots and that the connected component that contains the identity is given by the torus . to prove the second statement , we need to show that for every pair @xmath73 of distinct indices of rays ( for brevity , we will often write @xmath74 meaning a ray of a fan or its primitive generator @xmath75 ) , @xmath76 is different from @xmath77 inside @xmath78 . assume that @xmath73 is such a pair . then , by definition , @xmath67 and @xmath79 is a root . but we have shown that @xmath3 has no roots so the partition given by @xmath80 is the trivial one . in particular , @xmath81 is the trivial group . we have shown that , by changing the projective subdivision , one can change the part of @xmath25 that depends on the automorphism of the fan . this depends strongly on how much the subdivision is invariant under the hyperoctahedral group @xmath82 that is the group of the automorphisms of the hypercube and of the fan spanned by its faces . in particular , @xmath83 will be a subgroup of @xmath82 . an example of maximal projective subdivision that preserves the full symmetry of the hypercube is the one that has , as maximal cones , the cones spanned by a flag in the hypercube . by a _ flag _ , we denote a tetrahedron with one vertex on the center of one of the @xmath48 cubes of @xmath2 ( call it @xmath84 ) , another vertex on the center of a @xmath47face of @xmath84 ( denote the face by @xmath85 ) , the third vertex on an edge @xmath86 of @xmath85 and the final vertex on one of the two end - point of @xmath85 , i.e. on a vertex of the hypercube . the set of such flags cover the hypercube and two flags meets only along a lower dimensional face of the two tetrahedra . this implies that the set of the cone spanned by the flags generate a fan with @xmath87 facets ( one for each flag ) . by construction , the set of the primitive generators of each maximal cone - which are the vertices of a flag in the hypercube - is a @xmath88basis for @xmath89 . this imply that the resolution is smooth and is a maximal subdivision of @xmath39 . it is also easy to see that the corresponding toric variety is projective by constructing an explicit support function iteratively . we will call this subdivision the _ flag subdivision _ of @xmath4 . maximal cones . in the picture we show the slice of the subdivision of the cone corresponding to the cube with center @xmath90 with the plane @xmath91 . on the left , one of its @xmath92 tetrahedra ( each tetrahedron corresponds to one of the maximal cones of the flag subdivision ) . ] to compute the invariants of the action of a subgroup of @xmath25 on the cohomology of @xmath3 , it is sufficient to consider the action of its projection on @xmath83 ; in fact , @xmath12 acts trivially . let @xmath93 be a subgroup of @xmath82 and assume that a @xmath94invariant subdivision has been chosen . call @xmath7 the generic element of @xmath95 that is a smooth calabi - yau that it is the mirror of @xmath6 . assume moreover that @xmath7 is @xmath96invariant with @xmath97 such that the image of @xmath98 in @xmath83 is @xmath93 . since @xmath3 is simplicial ( i.e. , it has only orbifold singularities ) , the picard group has finite index in @xmath78 . this implies that @xmath99 . consider the exact sequence @xmath100 & { \mathbb{z}}^4 \ar[r]^-{\alpha } & { \mathbb{z}}^{{\tilde{\sigma}}(1)}:=\bigoplus_{\rho\in { \tilde{\sigma } } } { \mathbb{z}}\ar[r]^-{\beta } & { a}_{3}({\tilde{x } } ) \ar[r ] & 0}\ ] ] where @xmath101 and @xmath102 . by tensorizing with @xmath103 , we obtain the exact sequence @xmath104 & { \mathbb{q}}^4 \ar[r]^-{\alpha } & { \mathbb{q}}^{80 } \ar[r]^-{\beta } & \operatorname{pic}_{{\mathbb{q}}}({\tilde{x } } ) \ar[r ] & 0},\ ] ] from which we deduce that @xmath105 has dimension @xmath106 . since @xmath107 , there is a map from @xmath105 to @xmath108 . by standard facts in toric geometry ( see , for instance , @xcite ) , the kernel of this map is generated by the classes @xmath109 of divisors associated to edges of @xmath110 spanned by integral vectors , which are in the interior of some facets of @xmath2 . in our case , @xmath2 is the hypercube and the toric divisors in the kernel are the ones corresponding to the centers of the @xmath48 cubes of @xmath2 . therefore , the dimension of the kernel in @xmath48 . consider the exact sequence @xmath111 & \operatorname{ker}\ar[r ] & \operatorname{pic}_{{\mathbb{q}}}({\tilde{x } } ) \ar[r ] & \operatorname{pic}_{{\mathbb{q}}}({\tilde{y } } ) \ar[r ] & 0}.\ ] ] if we take the invariant subspaces in ( [ eq : exseq ] ) , we obtain the exact sequence ( see , for instance , @xcite ) @xmath112 & \operatorname{ker}^k \ar[r ] & \operatorname{pic}_{{\mathbb{q}}}({\tilde{x}})^k \ar[r ] & \operatorname{pic}_{{\mathbb{q}}}({\tilde{y}})^k \ar[r ] & 0}\ ] ] that allows us to calculate the invariant part of @xmath108 . since @xmath7 is a smooth calabi - yau , we have @xmath113 . we then need to search for a group @xmath93 of order @xmath24 such that the dimension of @xmath114 is @xmath115 . up to conjugacy , the group @xmath82 has exactly @xmath116 subgroups of order @xmath24 . by means of a computer search , we can compute the dimension of the vector spaces in the previous exact sequence . the following table summarizes these results . for each group we specify also the gap index ( except @xmath117 ) . for example , groups of order @xmath24 are identified by a pair @xmath118 where @xmath119 is a progressive number . the gap index of a group of order @xmath24 is @xmath119 . ] by which it is possible to identify the isomorphism class of the group from the gap database . for example , the group relative to line @xmath120 of the table , has gap index @xmath121 that corresponds to @xmath122 . [ cols="^,^,^,^,^,^,^",options="header " , ] as it is possible to see from the table , there is only one group ( up to conjugacy ) such that @xmath123 . this group has gap index @xmath23 and it is isomorphic to a particular semidirect product of @xmath124 and @xmath125 , which is denoted by @xmath126 with the following presentation @xmath127 by direct inspection , one checks that @xmath82 has exactly @xmath23 subgroups isomorphic to @xmath126 and they are all conjugated . then the following result holds . assume that @xmath42 is a maximal projective subdivision of @xmath33 such that @xmath128 admits a smooth anticanonical section @xmath7 and a group @xmath98 with @xmath129 . then , @xmath42 must have a @xmath130symmetry . in what follows , @xmath42 is a maximal projective subdivision of @xmath39 that has a @xmath130symmetry . without loss of generality , we can assume @xmath131 with @xmath132 it will be useful to observe that @xmath126 has three involutions , namely @xmath133 , and that they generate a subgroup of order @xmath31isomorphic to @xmath134 . moreover , among the @xmath41 involutions , only @xmath135 has square roots ( that are @xmath136 and @xmath137 ) . recall that we need a group @xmath98 of order @xmath24 such that its projection on @xmath83 is @xmath93 . now we will describe all such groups @xmath98 . as before , call @xmath12 the maximal torus inside @xmath3 . recall that @xmath12 is an open dense subset of @xmath3 and that we can write it as the toric affine patch corresponding to the cone @xmath138 od @xmath42 , i.e. , it has a realization inside @xmath3 given by @xmath139)={spec}({\mathbb{c}}[x_1^{\pm 1},x_2^{\pm 1},x_3^{\pm 1},x_4^{\pm 1}]).\ ] ] choose coordinates @xmath140 on @xmath12 such that the action of the torus on @xmath141 is given by @xmath142 in coordinates , for every @xmath143 one has @xmath144 the action of @xmath93 on the points of @xmath145 can be deduced from that on the coordinates . for example , consider @xmath146 whose action on the lattice @xmath147 is described by the matrix in ( [ eq : gropgens ] ) . the action of its dual @xmath148 is described by the transpose of the matrix in ( [ eq : gropgens ] ) : it is a permutation on the set @xmath149 . here , we denote by @xmath150 the elements of the standard @xmath88basis of @xmath61 and @xmath151 its dual . now , recall that the coordinates @xmath152 on the torus are , by definition , of the form @xmath153 . this allows us to deduce the action of @xmath154 ( as of any other element of @xmath83 ) on the torus . thus the description of the action of @xmath154 and @xmath155 on @xmath12 ( which will be denoted again by @xmath154 and @xmath155 ) are @xmath156 @xmath157 two generators of the group that projects on @xmath93 are of the form @xmath158 , where @xmath159 . by direct inspection , one can check that @xmath160 is isomorphic to @xmath126 if and only if the following conditions are satisfied : @xmath161 that is to say , @xmath162 thus , we have @xmath163 families ( one for each choice of @xmath164 ) of groups the satisfies our requirements . if one takes @xmath165 to be @xmath166 with @xmath167 then @xmath168 @xmath169 with @xmath170 so , up to conjugacy , there are only two subgroups of @xmath25 that project themselves on @xmath93 : @xmath171 define @xmath172 and @xmath173 , so that @xmath174 . in this section , we analyse invariant sections and their relation with the fixed locus of @xmath176 and that of @xmath177 . for most of the section , the toric variety is a generic maximal projective resolution @xmath3 that has the @xmath130symmetry described before . for more details about fixed points we will , however , restrict to consider the flag resolution defined at the end of section ( [ sec : autores ] ) . first , we will analyse the invariant space of anticanonical sections of @xmath176 and @xmath177 . by standard toric geometry , @xmath178 has dimension equal to @xmath179 . moreover , from this set it is possible to write down explicitely a basis for the space in the coordinates of the torus @xmath145 ( which we will identify with @xmath12 ) . more precisely , given a divisor @xmath180 in a projective toric variety @xmath0 constructed from a polytope in a lattice @xmath181 , there is an isomorphism between @xmath182 and the set @xmath183 where @xmath184 is the polytope associated to the divisor @xmath180 , @xmath185 are the coordinates of the torus and @xmath186 is a shorthand for @xmath187 . in our case , because @xmath42 is a maximal projective subdivision , one has @xmath188 , the @xmath34cells . the integral points of @xmath5 , i.e. the points of the set @xmath189 , are the coordinate points , their opposite and the origin . hence there are @xmath190 indipendent anticanonical sections and @xmath191 from this description , it is also easy to describe the invariant subspaces with respect to the actions of @xmath176 and @xmath177 . we will write @xmath192 to denote a subspace of @xmath178 such that @xmath193 and @xmath194 for every @xmath195 . by direct computation , we obtain @xmath196 @xmath197 where @xmath198 is a subspace that does nt contains any invariant subspace of dimension @xmath13 . apart from @xmath199 which has dimension @xmath163 , every other subspace of the form @xmath192 has dimension @xmath13 . the invariant subspace with respect of @xmath176 has dimension @xmath163 and it is generated by @xmath200 the generic invariant section is then @xmath201 whose zero locus can be proven to be smooth in @xmath12 for generic values of @xmath202 and @xmath203 . its closure inside @xmath3 is an invariant anticanonical section . let s focus on the fixed locus of @xmath176 on @xmath12 . denote by @xmath204 the subgroup of @xmath176 generated by @xmath135 and @xmath205 . as observed before , it is isomorphic to @xmath134 and consists of all the involutions of @xmath176 plus the identity . for this reason , the fixed locus of @xmath176 is that of @xmath204 . in @xmath93 , @xmath135 equals to @xmath206 , so @xmath135 acts on @xmath12 by sending @xmath185 to @xmath207 . the fixed locus of @xmath135 inside @xmath12 consists of those points on the torus with coordinates @xmath185 such that @xmath208 i.e. , the @xmath24 points with @xmath209 . by direct inspection , the generic section @xmath210 does nt vanish on any of these points . the fixed points of @xmath205 are those for which @xmath211 so they have @xmath212 and @xmath213 free . every section meets each of the four component of the fixed locus in a curve that is smooth on @xmath12 . a similar result is true for @xmath214 . this shows that the generic @xmath215invariant section meets the fixed locus of @xmath176 in @xmath48 curves and is smooth in @xmath12 . the elements of @xmath176 that have fixed points ( on @xmath12 ) on @xmath216 are only @xmath155 and @xmath217 because @xmath135 is the only involution that has a square root and it acts freely . it remains to consider what happens outside @xmath12 , i.e. , on the toric divisors of @xmath3 . [ lem : fixpint ] let @xmath218 be a fixed point of @xmath154 in @xmath3 . then @xmath218 lies in @xmath12 . if @xmath218 is a fixed point of @xmath154 , then it is also a fixed point for @xmath135 . the toric variety @xmath128 is covered by affine patches @xmath219 where @xmath220 is a facet in @xmath42 . in particular , there exists @xmath220 such that @xmath221 . call @xmath222 the cone that is the image of @xmath220 by the involution @xmath135 . being a fixed point , @xmath218 will also be a point of @xmath223 because @xmath224 . we have then @xmath225 but @xmath135 acts on @xmath89 as @xmath206 so @xmath226 . being @xmath220 a strictly convex rational cone one has then @xmath227 . in the end , hence @xmath228 . for every maximal projective subdivision @xmath3 of @xmath4 that admits a @xmath94symmetry , the automorphism @xmath154 acts freely on the whole generic invariant anticanonical section . assume now that @xmath3 is the flag subdivision . let s prove that the closure of the generic @xmath216 is smooth in the affine patch @xmath219 where @xmath220 is the facet generated by the flag whose vertices are @xmath229 the dual cone is generated by the following primitive vectors @xmath230 so that the @xmath231algebra generated by the semigroup @xmath232 is @xmath233={\mathbb{c}}[x_1,x_2,x_3,x_4].\ ] ] the affine coordinates on @xmath219 are then @xmath234 . the relations that allow to go from the @xmath235 to the coordinates on the torus are @xmath236 making the substitution one has @xmath237 we multiply by @xmath238 because , in this way , one does nt change the zero locus in @xmath12 ( in fact @xmath238 is invertible in the coordinate ring of @xmath12 ) but allows , with the substitution , to cancel other components that are not in the closure . by direct inspection one sees that @xmath50 is ( for generic @xmath239 ) smooth . this shows that there exist a smooth calabi - yau threefold @xmath7 that is @xmath215invariant ( and that is also a mirror for @xmath6 ) . as a consequence of a result is a calabi - yau threefold and @xmath240 is a small involution acting on @xmath0 such that its fixed locus contains a curve , then @xmath240 is a symplectic automorphism of @xmath0 ] in @xcite , we deduce that @xmath176 is a symplectic subgroup of @xmath242 . in fact @xmath154 acts freely on @xmath7 ( by direct check inside the torus and using the lemma for the outside ) and @xmath155 is an involution that fixes curves . with this we are able to complete the proof of the following theorem . [ thm : mirror ] there exists a maximal projective subdivision @xmath42 ( i.e. the flag subdivision ) of @xmath39 and its associated toric variety @xmath3 , a smooth calabi - yau threefold @xmath7 in @xmath3 and a group @xmath97 ( namely @xmath176 ) such that : * @xmath7 is @xmath96invariant and a mirror for @xmath243 ; * @xmath98 is symplectic as subgroup of @xmath242 and acts with a fixed locus with irreducible components that are curves ; * @xmath244 is a calabi - yau orbifold whose singular locus has pure dimension @xmath13 and it has fundamental group isomorphic to @xmath245 ; * @xmath8 is a topological mirror for @xmath246 , i.e. it has @xmath247 . the only thing left is that @xmath8 is a topological mirror for @xmath246 , i.e. we have to prove that @xmath248 and that its fundamental group is @xmath245 . following batyrev s work we have that the space @xmath249 classifies the infinitesimal deformation of @xmath7 . it has a subspace @xmath250 which parametrizes the deformation determined by the hypersurfaces @xmath251 . the other , possible , deformations of @xmath7 form a subspace of dimension @xmath252 where @xmath253 runs over all the faces of @xmath5 of codimension @xmath163 and @xmath254 is the face of @xmath2 dual to @xmath253 . but @xmath5 is the @xmath34cells and each face has no points in the relative interior so @xmath255 . this shows that all the contribution to @xmath256 cames from @xmath257 . hence , it suffices to analyse the action of @xmath176 on @xmath258 . but , as we have seen before , @xmath259 and the action of @xmath176 on this space is determined by the permutation of the points in @xmath5 . we have already calulated the invariant subspace and it has dimension @xmath163 . this implies that @xmath260 . in ( @xcite ) , it has been calculated the fundamental group of the generic anticanonical section of a fano fourfold . only a few of the varieties analysed have non - trivial fundalmental group and the case which concern us now , the anticanonical section in the dual of @xmath1 , it is not one of those . this shows that @xmath261 is simply connected and thus the same holds for @xmath7 . by the main result of ( @xcite ) , the fundamental group of @xmath7 is isomorphic to the quotient of @xmath98 by the group generated by the elements whose fixed locus is not empty . as we have seen the only automorphisms with fixed points on @xmath7 are @xmath205 and @xmath214 so the fundamental group of @xmath8 is @xmath262 . in this case , the decomposition of @xmath178 does nt have invariant spaces of dimension greater than one that are intersections of eigenspaces of elements of @xmath177 . so , to each invariant subspace of dimension @xmath13 ( there are @xmath115 of those ) we associate an anticanonical section that is @xmath263invariant . we will see in a moment that none of these is a mirror for @xmath6 : they are all singular . consider the subspace @xmath264 . this is spanned by the section @xmath265 so its zero locus on the torus is empty . the zero locus of the corresponding section on @xmath3 is the union of the toric divisors ; hence it is not only highly singular , but also reducible . the other invariant subspaces are of the form @xmath266 for @xmath267 . if we denote by @xmath268 an element that spans @xmath266 , we can consider the following sections : @xmath269 @xmath270 @xmath271 @xmath272 one can show that the points of the torus with coordinates @xmath273 @xmath274 are the singular points of @xmath275 whereas @xmath276 are the singular points of @xmath277 for @xmath278 . this concludes the case of @xmath177 . we have also proved a refined version of theorem ( [ thm : mirror ] ) : indeed , we have also shown that if the pair @xmath279 exists for some @xmath3 , as in the hypotesis of theorem [ thm : mirror ] , then @xmath98 is uniquely determined by its conjugacy class . by theorem ( [ thm : mirror ] ) , we have seen that there exists a topological mirror for @xmath246 and it is constructed as @xmath280 for suitable @xmath7 and @xmath98 . furthermore , @xmath8 is singular and all its deformations have at least the same singularities as @xmath8 : we have an equi - singular family of anticanonical quotients that is the ( topological ) mirror - family for @xmath246 . there is no hope for a smoothing to obtain a smooth mirror for @xmath246 from @xmath8 . one can instead consider a crepant resolution . this exists because @xmath7 is a smooth calabi - yau threefold and @xmath98 is a symplectic group . the crepant resolution is again a calabi - yau , it is smooth but it will have different hodge numbers that are given by the dimension of the orbifold cohomology of @xmath281 . now , we will analyse the fixed locus of the group @xmath176 action on the toric resolution associated to the flag subdivision and compute the orbifold cohomology for the quotient . we recall that the fixed locus of @xmath176 on the torus @xmath12 is given by the union of the fixed loci of @xmath205 and @xmath214 . these two elements are the only ones that have non - empty fixed locus ( on @xmath7 ) . both are symplectic involutions of @xmath7 and have fixed locus smooth of pure dimension @xmath13 . if a curve of @xmath282 does nt intersect @xmath12 then it is contained in @xmath283 i.e. , in the union of the toric divisors ( restricted on @xmath7 ) . now , we will show that there are no more curves other than those we have found in @xmath12 . assume that @xmath218 is a fixed point of @xmath205 . the same methods apply also to @xmath214 . call @xmath220 a maximal cone of @xmath42 such that @xmath284 . as we have done in lemma ( [ lem : fixpint ] ) , if @xmath285 we have @xmath286 in this case , however , it is indeed possible that @xmath287 and these are the cases we have to investigate . all the maximal cones of the flag subdivision have primitive generators for their rays of the form @xmath288 where @xmath289 . to have @xmath290 it is necessary that @xmath291 is fixed by @xmath205 . this happens if and only if @xmath292 . if that is the case , we have two further cases depending on whether @xmath293 is in @xmath294 or not . if @xmath295 then @xmath296 is the cone generated by @xmath297 and @xmath298 . in the other case @xmath296 is a ray and has , as primitive generator , the vector @xmath297 . , a tetrahedron and its image with respect to @xmath205 in two cases : the right one when @xmath299 and the second when @xmath300 . ] assume that @xmath301 . in this case @xmath302 and @xmath303={\mathbb{c}}[x_1,x_2^{\pm 1},x_3^{\pm 1},x_4^{\pm 1}].\ ] ] the conditions to be satisfied by the coordinates of a fixed point of @xmath205 inside @xmath304 are @xmath305 . if we define @xmath306 and @xmath307 such that @xmath308 for @xmath309 , the equations for the fixed locus of @xmath205 on @xmath7 ( restricted to @xmath304 ) are @xmath310 for each choice of @xmath311 we have , generically , a smooth and irreducible curve that is not contained in the locus @xmath312 . our interest in this fact is justified because the points of @xmath304 with @xmath313 are exactly the points of the torus . if we want to find new components for @xmath282 , we have to search in the locus @xmath312 . we have then show that there are no other components of @xmath282 in @xmath304 . the projection of @xmath176 on @xmath314 acts transitively on @xmath315 so this shows that there are no more components of the fixed locus on @xmath316 it remains to consider the case @xmath295 . in total there are @xmath48 cones in @xmath42 with these properties and they are arranged in @xmath163 orbits under the action of @xmath126 . the orbits representatives are @xmath317 assume that @xmath318 . the dual cone is @xmath319 so that @xmath320={\mathbb{c}}[x_1,x_2^{\pm 1},x_3^{\pm 1},x_4x_1^{-1}]:={\mathbb{c}}[w_1,w_2^{\pm 1},w_3^{\pm 1},w_4].\ ] ] in this affine patch , @xmath7 can be described as the zero locus of @xmath321 the action of @xmath205 on these coordinates is given by @xmath322 so the fixed locus is given by @xmath323 . these conditions give @xmath31 irreducible curves that are the closure of the ones we have found in the torus . the same argument applies when one search for new components in @xmath324 . in conclusion , @xmath205 has a fixed locus composed of @xmath31 irreducible curves and none of them is contained in a toric divisor . the only other element that has fixed points in @xmath176 is @xmath214 but this is conjugated in @xmath176 to @xmath205 . this means that none of the components of the fixed locus of @xmath176 is containtd in the intersection of @xmath7 and a toric divisor . it is also interesting to note that , by the main result of ( @xcite ) , the fundamental group of @xmath7 is @xmath325 . [ thm : resolution ] a crepant resolution of @xmath8 is a smooth calabi - yau manifold with hodge numbers @xmath27 . the formula for the orbifold cohomology of @xmath280 is the following : @xmath326 where @xmath327 is a set whose elements are representant for the conjugacy classes of @xmath98 , @xmath328 is the centralizer of @xmath329 in @xmath98 . in our particular case we can simplify the formula a lot because the only elements with fixed points are @xmath205 and @xmath214 and they are conjugated . being @xmath205 a symplectic involution one has that its age is @xmath13 so the formula becomes @xmath330 the shifts on the degree mean that we have a contribution to @xmath331 equal to the number of the components of @xmath332 and a contribution to @xmath333 equal to the sum of the genus of the components of @xmath332 . the centralizer of @xmath205 is the group generated by @xmath136 and @xmath155 that is abelian and isomorphic to @xmath334 . the action of @xmath335 is not faithful ; in fact , @xmath205 acts trivially . this means that @xmath336 . if , as before , @xmath210 is the section corresponding to @xmath7 , denote @xmath337 the four components of @xmath338 . by definition it is clear that they are pairwise disjoint . we need to understand how @xmath136 acts on them . the action of @xmath136 on the coordinates is given by @xmath339 so @xmath340 and @xmath341 . it is also of interest to note that @xmath342 . this means that @xmath343 and each of these quotients is free ( because @xmath135 , and hence @xmath136 , does nt have fixed points ) . by the riemann - hurwitz formula we have @xmath344 and @xmath345 this , in particular , implies that @xmath346 and @xmath347 are congruent to @xmath13 modulo @xmath31 and that @xmath348 is odd . the map @xmath349 is finite and it is ramified along @xmath48 curves . these are @xmath350 ( the fixed locus of @xmath205 ) and the images of these curves by @xmath154 ( the fixed locus of @xmath214 ) . the two fixed locus are interchenged by @xmath154 so the branch locus is given by @xmath332 . we have already seen that is composed of @xmath41 curves . by writing explicitly the riemann - hurwitz formula , we obtain the following relation : @xmath351 this , and the arithmetic relations satisfied by @xmath352 , allow us to conclude that each component of the fixed locus and of the branch locus is an elliptic curve . in conclusion we have @xmath353 to conclude it remains to recall that the orbifold cohomology of a global quotient of a calabi - yau threefold by a finite symplectic group is the cohomology of one of its crepant resolution ( cfr . @xcite ) . [ osette ] in @xcite , the authors studied a particular calabi - yau threefold . we will briefly recall its construction . first of all , denote by @xmath50 the complete intersection in @xmath354 of the following two quadrics : @xmath355 by simple calculation , @xmath50 is proved to be smooth and a del pezzo surface of degree @xmath31 . let @xmath356 and @xmath357 be the automorphisms of @xmath358 such that @xmath359 ^ -r & \left((x_0 : x_1 : -x_2 : x_3 : -x_4 ) , ( y_0 : y_1 : -y_2 : y_3 : -y_4)\right ) \\ ( x , y ) \ar@{|->}[r]^-t & \left((y_0 : y_1 : -y_2 : -y_3 : y_4 ) , ( x_0 : x_1 : x_2 : x_3 : x_4)\right ) . } \ ] ] the group @xmath360 acts on the product @xmath361 . moreover , the generic invariant section of the anticanonical bundle of @xmath362 is smooth and the fixed locus of @xmath15 is disjoint from it . hence , the zero locus @xmath12 of such a section is a smooth calabi - yau threefold and admits a free action of @xmath15 . call @xmath363 one of the cyclic subgroups of @xmath15 of order @xmath31 . @xmath363 acts freely on @xmath12 as well . our interest in such variety is that the quotient @xmath364 is a smooth calabi - yau with hodge numbers @xmath365 . this implies that the calabi - yau @xmath8 of theorem ( [ thm : resolution ] ) yields an explicit example of a hodge - theoretic mirror for @xmath366 . both these varieties have the same fundamental group . in the previous section we have seen that if we start from a maximal admissible pair @xmath9 in @xmath0 , we are not able to find an admissible pair @xmath279 in any @xmath3 obtained from a maximal projective subdivision of @xmath39 such that @xmath20 and @xmath367 are hodge - theoretic mirror calabi - yau . if we , instead , start from a non - maximal admissible pair this is possible in some cases . we recall that in ( @xcite ) it is shown that two groups that are part of two admissible pairs in @xmath0 are isomorphic if and only if they are conjugated in @xmath368 . so there is no ambiguity by writing @xmath369 . it is also shown that the quotients associated to the admissible pairs @xmath370 and @xmath371 in @xmath0 have respectively hodge numbers @xmath372 , @xmath373 and @xmath374 . for these quotients , each fundamental group is isomorphic to the group of the admissible pair . we have all the information to prove the following theorem . [ thm : mirror2 ] let @xmath3 denote the flag - resolution of @xmath4 and let @xmath7 be a smooth calabi - yau that is @xmath375invariant ( such is the one used in theorem ( [ thm : mirror ] ) ) . then @xmath376 , @xmath377 and @xmath378 are admissible pairs in @xmath3 whose associated quotients satisfy : in particular , @xmath385 and @xmath386 are respectively hodge - theoretic mirrors for the quotients associated to the admissible pairs @xmath370 and @xmath371 in @xmath0 . each pairs of mirrors has the same fundamental group . we have already shown that @xmath7 is a smooth calabi - yau threefold which is invariant with respect to @xmath387 and that @xmath154 , an automorphism of order @xmath48 , acts freely on @xmath7 so that @xmath388 is an admissible pair in @xmath3 for @xmath389 . to obtain @xmath390 we adapt the same strategy as that used in section ( [ sec : chosing ] ) . since the action of the group is free , we get @xmath391 and @xmath392 . as noted in the proof of theorem ( [ thm : mirror ] ) , @xmath7 is simply connected . the claim about the fundalmental group follows from the fact that the quotients are free .

let @xmath0 be the toric variety @xmath1 associated with its four - dimensional polytope @xmath2 . denote by @xmath3 the resolution of the singular fano @xmath4 variety associated with the dual polytope @xmath5 . generically , anticanonical sections @xmath6 of @xmath0 and anticanonical sections @xmath7 of @xmath3 are mirror partners in the sense of batyrev . our main result is the following : the hodge - theoretic mirror of the quotient @xmath8 associated to a maximal admissible pair @xmath9 in @xmath0 is not a quotient @xmath10 associated to an admissible pair in @xmath3 . nevertheless , it is possible to construct a mirror orbifold for @xmath8 by means of a quotient of a suitable @xmath7 . its crepant resolution is a calabi - yau threeefold with hodge numbers @xmath11 . instead , if we start from a non - maximal admissible pair , in same cases , its mirror is the quotient associated to an admissible pair .

the near - infrared color excess of extincted background stars can reveal the dust distribution in a dense core and provide information on physical structure . historically , this technique has been hampered in small , dense regions by limited instrumental sensitivity ( jones et al . 1980 , 1984 ) , but the development of large format array cameras has sparked renewed interest ( lada et al . 1994 , alves et al . 1998 , lada , alves & lada 1999 , alves , lada & lada 1999 , 2001 ) . the unprecedented sensitivity available with the near infrared camera and multi object spectrometer ( nicmos ) instrument on the hubble space telescope ( hst ) opens up a new regime . in particular , the high extinctions of dense cores that form the immediate environs of candidate low mass protostars are now accessible . molecular line surveys of nearby dark clouds have identified a large number of low mass dense cores ( myers , linke & benson 1983 ) of which roughly half are associated with young stellar objects detected by _ iras _ ( beichman et al . 1986 ) . these observations have contributed to the general paradigm that at least some low - mass stars like the sun form by the gravitational collapse of isolated , dense cores . in the standard theory , originating with shu ( 1977 ) , the dynamical collapse to a star / disk system occurs from the inside - out . first , an @xmath4 distribution is established prior to collapse as the cloud core loses magnetic and turbulent support through ambipolar diffusion and relaxes to a balance between gravity and thermal pressure . collapse is initiated at the center where the density is highest , and a wave of infall propagates outward at the sound speed . conditions inside the infall radius asymptotically approach free - fall , with @xmath5 and @xmath6 . modifications to the simple theory have been made to account for more realistic initial conditions , including slow rotation ( tereby , shu & cassen 1984 ) and residual magnetic fields ( galli & shu 1993 , li & shu 1997 ) . a similar radial density profile is also produced by the competing self - similar collapse theory of larson ( 1969 ) and penston ( 1969 ) . however , the larson - penston flow predicts far larger infall velocities than the shu theory , and has proved unsuccessful at explaining molecular line observations ( see e.g. zhou 1992 ) . unambigous identification of inside - out collapse motions has proved difficult for several reasons . first , the fact that collapse initiates in the center of a core means that it can be detected only with observations of sufficiently high angular resolution . further difficulty comes from the fact that the velocities associated with infall motions are small , comparable to intrinsic line widths and much smaller than the velocities of the bipolar outflows powered by young stellar objects . the presence of localized , redshifted self - absorption in spectral lines of moderate optical depth provides one signature of infall motions , but this signature is not unique . scenarios other than collapse can produce the same spectral line profiles . the most promising claims for collapse have been inferred from redshifted self - absorption , and consequently these claims are often controversial ( see the review of myers , evans & ohashi 2000 ) . the dense core in the b335 region is generally recognized as the best protostellar collapse candidate . the b335 region contains an isolated , roughly spherical globule at a distance of roughly 250 pc ( tomita , saito & ohtani 1979 ) . deeply embedded in the globule is a young stellar object discovered at far - infrared wavelengths by keene et al.(1983 ) and detected only at @xmath7 m by _ iras_. observations of high velocity molecular line emission have shown that the young stellar object drives a bipolar outflow with semi - opening angle of @xmath8 ( cabrit , goldsmith & snell 1988 , hirano et al . 1988 , 1992 ) . detailed radiative transfer models based on the theory of inside - out collapse provide good fits to many molecular line profiles of dense gas tracers observed at @xmath9@xmath10 resolution ( zhou et al . 1993 , choi et al . these models imply an infall radius of 0.03 pc ( @xmath11 ) , an age of @xmath12 yrs , and a central mass of @xmath13 m@xmath14 . while the agreement of the model with few parameters and the observed molecular line profiles strongly support inside - out collapse , the model is not unique . for example , in one extreme scenario discussed by zhou et al . ( 1993 ) , an unrelated foreground cloud of just the right excitation properties absorbs emission from a static core and exactly reproduces the infall signatures . recent molecular line observations of b335 made with the higher angular resolution of interferometry are also starting to show perhaps important discrepancies with the predictions of inside - out collapse ( wilner et al . 2000 ) . an extinction map of sufficient depth can probe the density structure of b335 and test the claim for inside - out collapse in a way that is free from many of the problems that plague molecular line studies , including excitation effects and spatially varying abundances of the molecular tracers . b335 is fortuitously placed for an extinction study , located at galactic coordinates @xmath15 , @xmath16 , which is a direction with a very high density of background stars with uniform properties . previous optical extinction studies of b335 were not sensitive to the structure of the dense core anywhere close to where the molecular line data suggest collapse ( see bok & mccarthy 1974 , dickman 1978 , tomita et al . 1979 and frerking et al . 1987 ) . in this paper , we present a near - infrared extinction study of b335 using data from hst / nicmos , supplemented by data from the w.m . keck observatory . section [ sec : obs ] briefly reviews the technique and describes the observations and data reduction . section [ sec : results ] presents an analysis of the extinction data in the context of several physical models for the b335 density structure , and comparisons with the results of previous studies . simulations of the observations are also presented , using a monte carlo approach detailed in the appendix . section [ sec : conc ] summarizes the main conclusions . measurements of the extinction of stars background to a dense core provides estimates of column density along many pencil beams . these estimates can be used to determine the overall structure of the obscuring core . the extinction measurements are best done at infrared wavelengths , where absorption due to dust is much less than in the optical , thereby probing regions of higher ( visual ) extinction . the basic method is to measure the near - infrared color excess for each background star : @xmath17 where @xmath18 is the intrinsic color of the star . the color excess represents the differential optical depth between the two near - infrared filters , which is directly proportional to dust column density . the observed color excess may be converted to an equivalent visual extinction using the standard reddening law ( e.g. rieke & lebosfsky 1985 ) , and then converted to gas column density by adopting a gas - to - dust ratio ( e.g. bohlin et al . there are some uncertainties in both of these conversions . the gas - to - dust ratio in protostellar cores is the subject of some debate ; at low extinction there is some evidence to suggest that it is constant along different lines of sight in the galaxy ( bohlin et al . 1978 ) , but at high extinction it is unclear whether the gas - to - dust ratio remains constant . for example , if grain growth occurs in the high extinction regions as a result of molecular depletion and agglomeration , then one might expect lower extinction for a given column density . the reddening law is probably less problematic and nearly universal at near - infrared wavelengths , given the small grain sizes present in protostellar envelopes ( mathis 1990 ) . in practice , the color excess method does not require a knowledge of the intrinsic colors of the individual extincted stars . a statistical correction to their observed colors may be obtained empirically from the background stellar population , provided it is sufficiently spatially uniform . for stars of spectral types a0 to late m , in the main sequence or the giant loci , the intrinsic colors lie in the narrow range @xmath19 ( koornneef 1983 ) . therefore , at high extinctions , little error is introduced by simply assuming a mean background star color derived from nearby control fields . the b335 globule was imaged with the nic3 camera on hst ( cycle 7nicmos , go-7843 , 1998 june 23 ) using two broadband filters , f160w and f222 m , similar to the usual h band and k band filters that are matched to the 1.6 @xmath20 m and 2.2 @xmath20 m atmospheric windows . the nic3 camera has @xmath21 pixels with size @xmath22 giving a @xmath23 field of view . the observations were obtained during the second of two periods of a few weeks when the hst secondary mirror was adjusted to allow the nic3 camera to be in focus ( the `` nic3 campaigns '' ) . the observing program had two parts : ( 1 ) a deep @xmath24 mosaic of the central region of high extinction , taking 6 orbits , and ( 2 ) a radial strip of images extending about @xmath25 from the center , taking 1 orbit , to reach the background stellar population and provide a sample of stars to transform nicmos magnitudes to a standard photometric system . the length of the strip was the maximum allowed before suffering the heavy overhead of guide star reaquisition . ( the strip included the center of b335 as insurance in the event that the deeper observations were unsuccessful . ) figure [ fig : skysurv ] shows the observed fields overlayed on a wide field optical view of the region from the digital sky survey . we note that the nic3 orientation angle on the sky and stepping direction of the radial strip were given by the spacecraft orientation on the day of observation . advance specification of a particular orientation would have created an untenable scheduling constraint ; the orientation obtained was based solely on where the spacecraft happened to be during the period when the nic3 camera was in focus and b335 was accessible . the central region was observed with a standard nine - point spiral dither pattern with @xmath9 offsets , and a total exposure time of 896 seconds per dither . the resulting square mosaic covers @xmath26 . the strip images were taken in _ one - chop _ mode , an exposure time of 56 seconds per chop , with an offset of @xmath27 between chops , and roughly @xmath28 overlap between successive images . this resulted in a strip of coverage @xmath29 wide that extends @xmath30 from the mosaic center . the data were taken in the nicmos _ multiaccum _ mode , for which an initial readout of the detector array is followed by non - destructive readouts during the course of a single integration . the intermediate readouts can be used to identify and to remove the effects of cosmic ray hits and saturated pixels . the nicmos data were obtained from the hst archive in fits format both unprocessed and already processed through the stsci pipeline ( calnic a & b ) . initial inspection showed the stsci pipeline did not adequately deal with a variety of instrumental effects , and so we reprocessed the data in iraf using the nicreduce software developed by brian mcleod ( mcleod 1997 , lehar et al . the nicreduce package provides superior and more versatile handling of sky and dark current subtraction , as well as cosmic ray hits and other defects . the bottom rows of the nicmos images suffered from a small amount of vignetting . the effect was most pronounced in the f222 m images , where about @xmath31 were unusable . the overlap of images in the dither pattern and in successive images of the strip was sufficient to ensure that the vignetting resulted in only a small loss in sensitivity in the affected regions , and no gaps in the spatial coverage . the dithered images of the central region of b335 suffered an additional complication due to the presence of a very bright star ( h magnitude of 9.0 ) on the west side of the mosaic center . the enormous count rate from the star caused the image reduction software to produce a small depression in the continuum level of the quadrant of each image that contained the star . this deviation was fixed by adding a constant to the relevant quadrant so that for each image the median pixel value was uniform in all four . ghost impressions of the star in successive exposures were interactively masked in the affected images and were not included in the final mosaic . stars were identified by visual inspection of the images , as several attempts at automatic star finding were compromised by the effects of diffraction spikes and rings around the brighter stars . photometry on each star was performed with the suite of routines in the _ apphot _ package . a series of apertures from @xmath32 to @xmath33 in radius was used for each star . aperture corrections were then performed using a minimum of four isolated and bright stars from the relevant image . each aperture was used to calculate a magnitude and associated error as a check to identify any unusual behavior . the measured count rates were corrected to a @xmath34 radius aperture , and then multiplied by 1.075 as detailed in the nicmos photometric calibration cookbook ( available from stsci ) . the instrumental magnitudes were then transformed to the vega scale using constants for each filter provided in the hst data handbook ( version 4 , dec 1999 ) . for stars that were observed in a region of overlap between strip images , weighted mean magnitudes were calculated . for stars observed in the overlap region of the central mosaic and the first strip image , the value from the more sensitive central mosaic was taken . the complete nicmos photometry list contains 1026 stars detected with the f160w filter and 471 stars detected with the f222 m filter with magnitude error @xmath35 . the nicmos limiting magnitudes are spatially dependent and are depicted in figure [ fig : limplot ] . in the central 32@xmath36 , with the most integration time , the limiting magnitudes are h=24.8 and k=21.5 ( in the cit system ) . observations of b335 were made on 1998 july 10 using nirc ( matthews & soifer 1994 ) on the keck i telescope . these observations were made for two reasons : ( 1 ) to obtain a better study of the stellar background population around b335 , and ( 2 ) to transform the nicmos f160w and f222 m magnitudes to a standard system . the nirc has @xmath21 pixels with size @xmath37 , giving a @xmath38 field of view . each basic observation consisted of nine dithered 5 second integrations , in a @xmath24 pattern with @xmath27 offsets , which were mosaiced together . observations were made with j , h , and k filters . the program had three parts : ( 1 ) nine images of the central region of b335 , approximately mimicking the sub - images of the nicmos central mosaic , ( 2 ) five images leading away from the center of b335 , at each position of a nicmos strip image , and ( 3 ) four additional fields @xmath39 away from b335 , to study the characteristics of the stellar background population . these regions are marked on figure [ fig : skysurv ] . the data reduction and calibration was done in iraf following the standard dark / flat / sky subtraction procedures , using a running flat . this procedure works well when the background emission does not change rapidly . for some periods during the observations , the sky emission in the h band did change relatively quickly , and the background subtraction for observations during these periods is not perfect . the typical seeing during the night was @xmath33 , though there were periods when it was better and also much worse . the majority of stars in the keck images of b335 could be identified with reference to the nicmos data , since the keck images are all less deep than the nicmos equivalents and the field of view slightly smaller . for the keck fields not observed with nicmos , all stars were again identified by visual inspection . photometry was performed for all non - saturated stars using the _ apphot _ package . again , a series of increasing radial apertures was used for each star , and each aperture was used to calculate a magnitude . this approach allowed a confidence level in the measurement to be assigned to each star , based on whether the distribution of measurements from different apertures was consistent with the predicted uncertainty . if the spread of the various magnitudes was greater than 0.2 magnitudes , then the measurement was identified as low quality ( 18% of the stars detected at h and 17% of the stars detected at k ) . these low quality stars have point - spread - functions that differ from the norm , mainly from those periods of poor and varying seeing conditions . the instrumental magnitudes were transformed to the cit system through observations of the hst infrared standards sj9164 and sj9182 ( persson et al . 1998 ) and the ukirt standard fs35 ( casali & hawarden 1992 ) . the internal dispersion in the zero points measured for these stars was 0.02 magnitudes , and no atmospheric extinction correction could be discerned . the keck photometry list contains a total of 297 stars detected at h , and 340 stars detected at k , in the overlap region between the keck and nicmos datasets . the four keck fields @xmath39 away from b335 contain typically 100 detected stars per square arcminute with h@xmath40 . the four fields display uniform properties and no discernable reddening ; the star counts in each image are identical to within poisson errors , and the mean @xmath41 color of the fields are all in the range @xmath42 to @xmath43 , consistent with the intrinsic colors of giant and main sequence stars . a detailed study of the background stellar population is presented in section [ sec : background ] . the limiting magnitudes of the keck observations are approximately h=19.7 and k=19.6 . for extinction calculations , the nicmos f160w and f222 m magnitudes are transformed to standard h and k magnitudes . linear least squares fits ( not including the lower quality keck data ) give slopes of @xmath44 for ( h@xmath45f160w ) vs. f160w , @xmath46 for ( k@xmath45f222 m ) vs. f222 m , and @xmath47 for ( h@xmath45k ) vs. ( f160w@xmath45f222 m ) , where the formal uncertainties in the fitted slopes are @xmath48 . the formal uncertainties are underestimates of the true uncertainties since the magnitude comparison slopes are strongly biased by the few brightest stars , and the color comparison slope is biased by the few reddest stars . experimentation with subsets of the stars resulted in variations in the fitted slopes at the @xmath49 level . we conclude that these data provide no evidence for non - zero slopes or color terms in the transformation . we calculated constant offsets for the magnitude conversions for each filter , using a weighted mean of ( h@xmath45f160w ) and ( k@xmath45f222 m ) values . the result is h = f160w@xmath50 , and k = f222m@xmath51 , where the uncertainties represent the standard error in the mean . figure [ fig : coltrans ] shows the scatter plot of @xmath41 vs. ( f160w@xmath45f222 m ) with the derived color transformation superimposed . for the @xmath41 colors of interest , in the range 0 to 3 , the differences between the two photometric systems are remarkably small . support for the accuracy of these transformations comes from published ground - based photometry of the one very bright star observed by krgel et al . ( 1983 ) , who measured johnson h and k fluxes of @xmath52 mjy and @xmath53 mjy , which correspond to magnitudes of h@xmath54 and k@xmath55 , and color @xmath56 . these values compare very favorably with the transformed nicmos magnitudes of h@xmath57 and k@xmath58 , and color @xmath59 . note that systematic errors limit the accuracy of the nic3 photometry to a few percent , and are not reflected in the random uncertainties quoted here . to check the uniformity of the background stellar population , the starcounts and mean colors of the four keck off fields were compared with those for the two nicmos strip images furthest away from the center of b335 . the colors in the different images are consistent within the observed dispersions . the fifth nicmos strip field has a mean color and standard deviation of @xmath60 , @xmath61 , very similar to that found for the combined keck off - fields : @xmath62 , @xmath63 . the fourth nicmos strip field shows a slightly redder mean color ( @xmath64 ) and higher dispersion ( @xmath65 ) than the other fields ; most likely , this field is not far enough away from the center of b335 to be completely free of its extinction , and this field was excluded from our characterization of the background stellar population . figure [ fig : backcolor ] presents a histogram of the @xmath41 colors of the `` background '' stars in 0.1 magnitude bins ( excluding the keck stars with lower quality photometry ) . the @xmath41 mean color is 0.13 magnitudes , with standard deviation 0.16 magnitudes . also shown for comparison is a gaussian distribution with the same mean and standard deviation ( normalized so that the area under each curve is equal ) . a color - magnitude diagram constructed for this sample shows no sign of any important systematic trends , e.g. there is no tendency for fainter stars to be redder . the starcounts in the fifth nicmos strip field are indistinguishable from those in the keck off - fields , taking into account the different different size fields of view . figures [ fig : images ] & [ fig : strip ] show the nicmos f160w and f222 m images of the b335 region . the central mosaic images in figure [ fig : images ] may be compared to h band and k band images of a smaller part of the central region of b335 obtained over three nights at keck by hodapp ( 1998 ) using nirc . in the region of overlap , the f222 m image shows the same stars as the k band image , but the overall sensitivity of the nicmos image is better and the point - spread - function much sharper , leading to better separation of nearby stars . the f160w image shows many more stars than the corresponding h band image . in contrast with the ground - based data , essentially all stars detected with the f222 m filter are also detected with the f160w filter because of the low background and greater sensitivity of the space - borne nicmos at the shorter wavelength . figure [ fig : nicmos ] conveys much of the information from the central mosaic of nicmos observations . the left plot is a pseudo - image in that the two axes are spatial , the apparent brightness of a star determines its `` size '' , and the value of its @xmath41 color determines its `` color '' . the right - hand plot displays the radial dependence of the @xmath41 colors out to @xmath66 from the center . the origin in both plots is given by the position of the embedded young stellar object , as determined by interferometric observations of millimeter continuum emission ( wilner et al . 2000 ) ; the uncertainty in the registration of the nicmos images and the millimeter peak is estimated to be about @xmath34 . based on an inspection of the extinctions , the stars in the right hand plot have been marked according to their spatial location : those that lie in the north or south quadrants are marked in red , and those that lie in the east or west quadrants are marked in blue ( the quadrant boundaries are marked in the pseudo - image as dashed lines ) . defining @xmath67 to be the angle subtended between the vector to a star and the axis of the outflow ( which runs very close to east - west ) , this division may be expressed simply as @xmath68 or @xmath69 . the nicmos central mosaic image contains 239 stars detected at h , 121 stars detected at k , and 119 stars that are detected in both filters . the detected star closest to the embedded young stellar object is at @xmath70 radius . the reddest detected star is at @xmath71 radius and has an @xmath41 color of 4.0 , corresponding to over 60 magnitudes of equivalent visual extinction . the major features of figure [ fig : nicmos ] are : ( 1 ) a strong gradient in the @xmath41 colors as one approaches the protostar ; ( 2 ) fewer stars are detected close to the center , despite better sensitivity ; ( 3 ) reddening of the stars turns off into background noise beyond the @xmath72 ( 0.15 pc ) outer radius of dense gas determined from molecular line data ( zhou et al . 1990 ) ; ( 4 ) there is a marked asymmetry in the stellar colors associated with the bipolar outflow stars viewed within @xmath73 of the outflow axis have systematically lower colors ; ( 5 ) the dispersion of the @xmath41 colors at a given radius and in the quadrants away from the outflow is @xmath74 greater than that of the background population . we describe several theoretical models for the density distribution of b335 and evaluate the success of these models in light of the extinction data . the models considered are not meant to comprise an exhaustive list . they include the inside - out collapse models previously proposed for b335 , bonnor - ebert spheres , and simple power law descriptions . we also discuss briefly the effects of the bipolar outflow and turbulent motions in b335 . there is a substantial literature of theoretical work by frank shu and colleagues describing the density structure of collapsing dense cloud cores , starting from the initial condition of a singular isothermal sphere . taking parameters appropriate for the b335 region , we briefly describe the pre - collapse state , the spherical inside - out collapse , and the modifications to spherical collapse expected for slow initial rotation or weak magnetic fields . _ pre - collapse configuration _ : the initial state is that of a singular isothermal sphere ; the density falls off as @xmath75 , where @xmath76 is the radius , with the normalization of the density determined by the effective sound speed . for b335 , zhou et al . ( 1990 ) derive a value for the effective sound speed of @xmath77 km s@xmath78 from molecular line data , which corresponds to a kinetic temperature of 13 k. the static initial condition is therefore given by : @xmath79 where the conversion to a molecular hydrogen number density assumes 0.7 for the the mass fraction in hydrogen . _ spherically symmetric collapse _ : the simplest scenario is the spherically symmetric collapse ( shu 1977 ) . in this model , a spherical wave of collapse propagates outwards from the center at the effective sound speed . the radial distance that the wave has traveled , sometimes called the infall radius , is the only additional parameter that defines the density distribution . inside the infall radius , conditions approach free fall , with the density taking the asymptotic form @xmath80 . by reproducing essential features of several molecular line profiles with this model , zhou et al . ( 1993 ) claimed evidence for inside - out collapse in b335 . choi et al . ( 1995 ) made more detailed radiative transfer models to match the zhou et al . ( 1993 ) observations of h@xmath81co and cs molecules , determining the infall radius to be @xmath82 pc ( @xmath11 ) . note that this derived infall radius falls entirely within the deep central mosaic of nicmos coverage . _ rotation _ : the effect of slow rotation on the collapse has been solved using a perturbative approach from the spherically symmetric case ( terebey , shu & cassen 1984 ) . the result is a distortion to the shape of the collapse wave of order @xmath83 , where @xmath84 is the number of times that the cloud has rotated since collapse initiated ( this is also equivalent to the ratio of the infall radius to the centrifugal radius ) . for b335 , frerking et al . ( 1987 ) found an angular velocity of @xmath85 s@xmath86 through c@xmath87o observations . this indicates a value of @xmath88 , and a distortion in the wave by only @xmath89 an entirely negligible effect . the density distribution is more strongly distorted , with the effect largest near the center of collapse . for radii greater than a tenth of the infall radius ( i.e.@xmath90 ) already well within the innermost radius where stars are detected the correction is always less than of order @xmath91 , which is also completely negligible . such slow rotation also has an unimportant effect on the model line profiles towards the center of b335 ( zhou 1995 ) . we therefore do not include rotation in our models . _ magnetic fields _ : the effect of a weak magnetic field on the inside - out collapse of a cloud has been studied by galli & shu ( 1993 ) and li & shu ( 1997 ) . the collapse wavefront and the density distribution are again distorted , in a similar way to the rotational model . in the magnetic case , the parameter @xmath84 is given by the ratio of the infall radius to the magnetic radius : @xmath92 galli & shu ( 1993 ) showed that the field effects only a small perturbation to the spherically symmetric solution outside a radius of @xmath93 . figure [ fig : nicmos ] shows that variation in color between the two spatial populations of stars can be as large as a factor of two , well beyond the infall radius of @xmath11 derived by choi et al . ( 1995 ) . since this certainly can not be construed as a `` small perturbation '' at these radii , we can place a lower limit on the magnetic field that would be needed to explain the asymmetry in the context of this model . taking @xmath94 , we find that : @xmath95 such a minimum magnetic field strength is implausibly large . for example , crutcher & troland ( 2000 ) measured a line - of - sight magnetic field of b@xmath96 @xmath20 g in the starless protostellar core l1544 using the zeeman effect in lines of oh , and this is much higher than the typical upper limits found for other dark clouds ( crutcher et al.1993 ) . in this context , it therefore seems very unlikely that a magnetically modified inside - out collapse could be responsible for the large asymmetry observed in the b335 extinction data . bonnor - ebert models are pressure confined isothermal spheres , for which the solution remains physical at the origin ( ebert 1955 , bonnor 1956 ) . in common with the singular isothermal sphere , the initial condition for inside - out collapse , they are solutions of a modified lane - emden equation ( chandrasekhar 1967 ) : @xmath97 where @xmath98 is the dimensionless radius , and @xmath99 is a logarithmic density contrast , with @xmath100 the ( finite ) central density . unlike the singular solution , the bonnor - ebert solutions do not diverge at the origin . instead , the boundary conditions are that the function @xmath101 and its first derivative are zero at the origin . the above equation can be solved by division into two first order equations which can then be tackled simultaneously using numerical techniques ( in our case a 4th order runge - cutta method ) . there is a family of solutions characterized by a single parameter the dimensionless outer radius of the sphere , @xmath102 . for a given sound speed and a particular choice of the shape of the density curve ( i.e. @xmath102 ) , there is one additional degree of freedom , the physical scale of the model . this additional degree of freedom may be removed by implementing an additional constraint , for example by fixing the outer radius , the total mass , or the external pressure . configurations with dimensionless outer radius @xmath103 are unstable to gravitational collapse ( bonnor 1956 ) . the gravitational collapse of bonnor - ebert spheres has been studied numerically by foster & chevalier ( 1993 ) . they find that the flow asymptotically approaches the larson - penston solution at the origin at the time of and prior to the formation of a central core , but emphasize that these large early accretion rates last for a very short time . if the cloud is initially very centrally condensed ( i.e.@xmath104 ) the later stages of infall closely resemble shu s inside - out collapse of a singular isothermal sphere . recently , alves , lada & lada ( 2001 ) used the color excess method to study b68 , a starless core , with data from the eso vlt . the b68 system is much less centrally condensed than b335 , and has much lower column density at the center ( only about 30 magnitudes of equivalent visual extinction ) . the alves et al . ( 2001 ) extinction map of b68 suggests that the density structure is well described by a bonnor - ebert sphere with dimensionless outer radius slightly in excess of critical : @xmath105 . although the b335 core is at a much later evolutionary stage than b68 , with an embedded protostar , we consider the bonnor - ebert sphere since it describes b68 so successfully . the outflows from young stellar objects can have a substantial impact on their surroundings . a bipolar outflow in b335 was discovered in co by frerking & langer ( 1982 ) . later studies in various tracers by cabrit , goldsmith & snell ( 1988 ) , hirano et al . ( 1988 , 1992 ) , chandler & sargent ( 1993 ) , and wilner et al . ( 2000 ) show that on arcsecond to arcminute scales , the outflow is well collimated with a semi - opening angle of about @xmath106 , and with the axis lying very nearly in the plane of the sky . the outflow orientation is close to east - west , with position angle roughly @xmath107 in the coordinate system of the nicmos images . inside the region affected by the outflow , the density structure is modified as the core material is swept out by the flow , but the exact form of the density field associated with the outflow is not known . the bipolar outflow may be an important driver of turbulent motions . zhou et al . ( 1990 ) determined a turbulent component to the effective sound speed in b335 of @xmath108 km s@xmath78 . coupling of the velocity and density fields via the continuity equation suggests that the perturbations to the velocity field should induce density contrast perturbations , i.e. @xmath109^{1/2 } \simeq 0.2 $ ] . the resulting perturbations to the column densities ( and extinctions ) will depend on the size scale of the velocity and density field perturbations . the perturbations to the column density will be largest when the size scale is a factor of a few smaller than the dense core . the extinction map is consistent with this possibility since the observed colors of neighboring stars are generally very similar , yet there exists a @xmath110% dispersion for stars that are not neighbors but are located at similar radial distances from the protostar ( see figure [ fig : nicmos ] ) . the effect of the column density perturbations are inevitably reduced by integration along the line - of - sight , but the observed residual turbulence most likely contributes significantly to the dispersion in the observed colors . we evaluate the goodness of fit for these various model density distributions by calculating a reduced @xmath111 , defined as : @xmath112 where @xmath113 is the number of free parameters in the model being fitted , and the sum extends over a particular subset of @xmath114 stars taken from the total number detected in both nicmos filters . the values of the observed color excess @xmath115 and the uncertainty @xmath116 are calculated using the properties of the unreddened stellar background population ( see section [ sec : background ] , and figure [ fig : backcolor ] ) , assuming each star to have an intrinsic color of @xmath117 , with an uncertainty of @xmath118 : @xmath119 \sigma_{i}^{\ , \mathrm{nic } } & = & \sqrt{\sigma_i^2 + \sigma_{\ , \mathrm{bg}}^2 } \nonumber \\ & = & \sqrt{\sigma_i^2 + 0.16 ^ 2 } \ , \end{aligned}\ ] ] where @xmath120 is the uncertainty in the observed @xmath41 color of a given star . this procedure implicitly assumes that the photometric errors ( @xmath120 ) and the intrinsic @xmath41 colors both have gaussian distributions . though the latter condition is not strictly true , monte carlo trials show that the @xmath121 values calculated with the above recipe agree very closely with those produced when the intrinsic color of each star is instead chosen randomly from the gaussian distribution with the same mean and standard deviation as the sample of background colors . we conclude that the results of the fitting are not sensitive to this approximation . since the models are in general non - linear in the fitting parameters , we analyze the uncertainty in the best - fit model parameters using a monte carlo technique known as the _ bootstrap _ method ( press et al . 1992 ) . to describe the method , let us represent the dataset that is being used in the fitting by @xmath122 . the first step is to construct a new dataset of equal size @xmath123 , by randomly selecting @xmath114 times from the original dataset . this new sample is then analyzed in the same way as the original dataset , and the fitted model parameters recorded . this process is then repeated @xmath124 times , where @xmath124 is sufficiently large that the resulting distribution of best - fit model parameters is insensitive to its exact value ( for our models , typically @xmath125 ) . the standard deviation of this distribution provides an estimate for the uncertainty of the parameters that best fit the original dataset . table [ tab : fits ] summarizes the results from the various @xmath126 analyses and these results are discussed below . figure [ fig : colschem ] shows a schematic diagram of the spatial coverage of the nicmos observations relative to the important size scales in b335 . the asymmetry in the observed colors appears to be associated with the bipolar outflow . since we do not know from theory the effect of the outflow on the material distribution , we initially restrict attention to the quadrants that are located away from the flow , @xmath68 , where @xmath67 is the angle subtended between the vector to a star and the axis of the outflow . to avoid noise near the boundary of the globule , we also restrict the fitting region to @xmath127 . in fit a of table [ tab : fits ] , we use a power law model for the density distribution . in fit b we use the shu inside - out collapse model , and in fit c we use a bonnor - ebert sphere . the best fitting single power law model ( fit a ) to the dense core region unaffected by the outflow has an index of @xmath128 . this value is well matched to the @xmath129 envelope of the isothermal sphere that represents the initial condition of the inside - out theory . the inside - out collapse solution ( fit b ) , which is similar to a broken power law with fixed indices , provides a somewhat better description than can a single power law model with arbitrary index . remarkably , the fitted infall radius of @xmath130 ( @xmath131 pc ) agrees within uncertainties with those calculated by zhou et al . ( 1993 ) and choi et al . ( 1995 ) by modeling molecular line profiles . that the minimum @xmath121 value for these fits are near 3 instead of unity reflects the @xmath74 greater dispersion in the observed colors than in the intrinsic colors of the background stars . a bonnor - ebert sphere provides an equally good description of the density structure in b335 ( fit c ) . figure [ fig : fitcomp ] shows a comparison of the @xmath41 radial profiles predicted by the bonnor - ebert sphere in fit c , and the inside - out collapse model in fit b. the plot shows that the two models are practically indistinguishable over the range in radius where there are stars to fit ( @xmath70@xmath132 ) . at large radii , the bonnor - ebert solution has slope close to a power law index of 2.0 , and at smaller radii , the slope flattens . in effect , the best fit bonnor - ebert sphere is one that mimics the inside - out collapse solution over the range in radius for which there is data . the models diverge only at such small radii that extinction data of sufficient depth to discriminate between them is currently impossible to obtain . one discriminant may be sensitive , high resolution observations of dust emission , which is strongest at these small radii where extinction data is lacking . regardless , the value of @xmath133 for the best fit bonnor - ebert sphere is well in excess of the critical value of 6.5 , and it is therefore an unstable configuration that should undergo gravitational collapse similar to shu s inside - out solution ( foster & chevalier 1993 ) . the simplest model we imagine to simulate the effects of the bipolar outflow on b335 is to assume that the outflow has completely cleared out the dense material in a bipolar cone of constant opening angle . figure [ fig : outflow ] shows the radial behavior of the @xmath41 colors from simulated observations we have made of an infalling model for four different values of the model outflow semi - opening angle @xmath134 . ( the appendix describes how the simulations are constructed . ) stars viewed through the model outflow cone are marked in red , and those not viewed through the outflow cone are marked in blue . the density distribution for @xmath135 in each of the simulations is the infalling model that best fits the @xmath136 data beyond @xmath137 from the outflow axis ( fit b ) . as the model opening angle increases , the stars begin to separate into two distinct populations . two effects are evident as the angle increases : ( 1 ) the population of detected stars with reduced colors becomes larger , and ( 2 ) the separation in the colors of the two populations grows larger . manifest in the simulation with the largest opening angle ( @xmath138 ) is that the first effect is not only caused by the outflow cone occupying a larger fraction of the sky , but is also due to there being less column density for lines - of - sight passing through the cone , which allows a greater number of the background stars to be detected . comparison of the radial behavior of the @xmath41 colors in these simulations with the observations in figure [ fig : nicmos ] shows that these simple models successfully reproduce the major features . the extent of the spread between the colors of the two populations of stars in the nicmos data seems to be reproduced best by the simulations with semi - opening angles of @xmath139 and @xmath137 . these angles also roughly reproduce the observed number of stars with significantly reduced colors . the simple hollow cone model of the bipolar outflow allows the stars viewed closer to the outflow axis to be included in the @xmath126 analyses . the fact that the predicted @xmath41 colors are a sensitive function of the model outflow opening angle means that we are able to fit for the opening angle of the flow . in both fits d & e , we fit for the entire region within @xmath127 , with a model which has both the infall radius and outflow opening angle as free parameters , and with the color excess normalization for each model chosen to minimize the @xmath121 in the region beyond @xmath137 from the outflow axis ( as in fit b ) . the only difference between fit d and fit e is that fit e includes an extra @xmath140 dispersion term in @xmath141 to account for additional structure in the cloud , perhaps from the residual turbulent fluctuations . table [ tab : fits ] lists the best - fit parameters for fit d and fit e , and they are nearly identical . in both fits d & e , @xmath121 is essentially an independent function of the infall radius and the outflow opening angle . figure [ fig : chisurf ] shows a contour plot of the surface of @xmath121 for fit d. the minimized value of @xmath121 is 4.02 , with best fit values of the infall radius and outflow opening angle of @xmath142 ( @xmath143 ) and @xmath144 , respectively . for fit e , the contour plots and fitted parameters are essentially identical infall radius @xmath145 ( @xmath146 ) and outflow opening angle @xmath147 but the minimized @xmath121 is @xmath148 due to the extra dispersion . that the two fits agree so closely , particularly with regard to the infall radius , is interesting . the extra dispersion in fit e reduces the weight of the most reddened ( and most central ) stars , which inevitably are those that constrain the fitting of the infall radius . the composite model with infall and bipolar outflow provides a consistent description of the density structure of b335 . when the regions affected by the outflow are excluded from consideration , the inside - out collapse solution provided a good description of the data . modeling the outflow as a hollowed - out bipolar cone allows inclusion of the stars from the outflow region in the @xmath126 analysis . despite the obvious crudeness of this model for the outflow , the fits to the entire dataset ( fits d & e ) constrain both the infall radius and outflow opening angle , and the fitted infall radii are entirely consistent with our earlier determination ( fit b ) and with the values determined by molecular line studies the fitted opening angle of the outflow is somewhat larger than indicated in studies of high velocity co emission . examination of the co velocity structure ( e.g. hirano et al . 1988 , 1992 ) suggests that the flow is more collimated at higher velocities , and that the opening angles dervied from co must be lower limits since the low velocity outflow becomes confused with the ambient globule emission ( and is resolved out by interferometers ) . this effect may explain why the opening angle appears wider in extinction than in co emission . of course , the one parameter model fitted here simply may not provide an adequate description of the outflow geometry at this level of detail . figure [ fig : models ] presents a simulation of the models for fit d and fit e to afford another way to examine the success of these models , as well as the differences between them ( the appendix describes the detailed recipe for constructing these simulations ) . the model for fit d has an infall radius of @xmath10 and an outflow opening angle of @xmath149 ( the normalization of the color excess is matched to the data by the @xmath126 analysis ) . this model reproduces the gross structure of the nicmos data very successfully . however , the radial plot shows that the dispersion in the @xmath41 colors at a given radius in the simulation is clearly significantly less than in the observations . the model for fit e is nearly identical to fit d except that the color excess has been multiplied by a gaussian random variable with mean of unity and standard deviation 20% . the simulated data from the model for fit e , with this additional dispersion to mimic the effects of residual turbulence , compares much more favorably with the observations in figure [ fig : nicmos ] . so far , we have been concerned only with the _ shape _ of the density profile and have ignored the free scaling factor applied to the color excess predicted by the various models during the fitting . for both the inside - out collapse models and the bonnor - ebert spheres , the normalization of the density distribution is entirely determined by the effective sound speed , measured to be @xmath150 km s@xmath78 ( zhou et al . this measurement , together with the distance to the cloud , gas - to - dust ratio , and near - infrared reddening law , make a unique prediction for the color excess normalization in each model . as listed in table [ tab : fits ] , a scaling factor @xmath15135 is required to match the models with the observed color excess . figure [ fig : modelb ] shows a simulation of the spherically symmetric inside - out collapse model with the effective sound speed of @xmath150 km s@xmath86 , the standard distance of 250 pc ( tomita et al . 1979 ) , and the infall radius of @xmath11 derived by choi et al . the simulation demonstrates the extent to which the model underestimates the observed reddening and extinction in b335 . if this model were correct , then the nicmos observations would have been sufficiently deep to see right through the center of the cloud , and the number of stars detected would have been far greater than observed . this therefore leads to the question : if the inside - out collapse model or the bonnor - ebert sphere model is correct , then what explains the need for this large scaling factor between the predicted extinction and the measured values ? the first possibility to consider is that the effective sound speed in b335 has been underestimated . to account for the @xmath15235 by a larger kinetic temperature in the initial state seems implausible a temperature of @xmath1534060 k would be needed . multi - transition observations of nh@xmath154 show that the bulk of the dense gas in b335 has a kinetic temperature of 1012 k ( menten et al . 1984 ) , consistent with a balance between heating and cooling processes . to account for the factor with non - thermal motions also seems implausible , given the measured widths of molecular lines . a non - thermal contribution of @xmath1550.350.45 km s@xmath78 would be required , which is clearly at odds with a multitude of molecular line data . for example , zhou et al.(1990 ) used observations of h@xmath81co to determine a small turbulent contribution of @xmath156 km s@xmath78 to the thermal sound speed . similar results from a variety of different molecules are documented by menten et al . ( 1984 ) and frerking , langer & wilson ( 1987 ) . one might envision a form of non - thermal support that is not reflected in the observed molecular line - widths , perhaps because of a special viewing geometry . if magnetic support were primarily from a static field and not waves , and this field were oriented so that the motions were not observable along the line - of - sight , then it would be effectively undetectable . although such a special situation seems contrived and unlikely , the orientation of b335 is special in that the outflow lies close to the plane of the sky , which may make such a configuration more plausible . another possibility that affects the overall scaling lies in the outer boundary , i.e. where the cloud ends . zhou et al . ( 1990 ) calculated the mass of b335 using an outer radius of 0.15 pc based on their radio map and the 250 pc distance estimate . the molecular line profiles of dense gas tracers are not particularly sensitive to this assumption . we have adopted this value for the outer radius throughout the extinction modeling . indeed , the column densities are also insensitive to this assumption since the density falls off steeply with radius . for a star at the infall radius of the best - fitting inside - out collapse model ( fit b ) , doubling the outer radius of the cloud increases the predicted extinction by only 7.5% , which is an unimportant amount with respect to the derived scaling factors . an overestimate of the distance to b335 would also result in predicted column densities that are too small . this distance uncertainty may be the dominant source of systematic error in the study . the distance of 250 pc was adopted from optical measurements of stellar reddening versus distance by tomita et al . ( 1979 ) , and this distance is uncertain at the 50% level . if b335 were closer , then the extinction would be larger at the same angular radius . to account for the full scaling factor would require that the distance be @xmath15735 times closer , only about 5080 pc , which seems unreasonable . also , for a given level of observed extinction , the inferred mass of the globule is proportional to the square of its distance ( since the angle subtended by b335 is fixed ) . for the outer boundary assumption of zhou et al . ( 1990 ) , the mass inferred from our extinction data is @xmath158 the 14 m@xmath14 for a distance of 250 pc may be large for a small globule , but reducing the distance by the whole scaling factor lowers the mass to less than 1 m@xmath14 for the inside - out collapse model , and a little over 1 m@xmath14 for the bonnor - ebert sphere . this seems small given that the globule must be forming a star with a substantial fraction of a solar mass , and losing a lot of mass through outflow in the process . furthermore , in the context of the inside - out collapse model , a reduction in the distance also results in a smaller mass for the protostellar core , as well as a shorter time since the onset of collapse ( both are proportional to the distance for a given sound speed and infall radius ) . in short , a closer distance for b335 would ameliorate the discrepancy between the observed and predicted extinction , but it is not clear if this could account for the entire difference . a final uncertainty is the conversion from column density to color excess using the gas - to - dust ratio and reddening law . significant variations in the reddening law have not been observed , and they are unlikely to explain much of the observed discrepancy in the color excess ( although we briefly return to this point in the appendix ) . by contrast , the gas - to - dust ratio has been observed to vary in the galaxy , with variations of @xmath159 not uncommon along different lines of sight ( bohlin et al . if the gas - to - dust ratio in b335 were lower than the standard value for the interstellar medium , or the reddening law steeper , then this would cause us to underestimate the extinction . potentially , the full scaling factor could be accounted for by this uncertainty , or perhaps by some combination of a closer distance and a modified gas - to - dust ratio . for example , if the true distance to b335 were actually 100 pc , and the gas - to - dust ratio were about one half of the standard value ( i.e. @xmath160 ) , then these two factors would fully account for the discrepancy between observed extinction and the predictions of either model , and also give a reasonable mass for the globule of @xmath161 . if molecular hydrogen were to deplete onto dust grains , this could cause such a decrease in the gas - to - dust ratio at high density . however , h@xmath81 is not thought to undergo significant freeze - out in this way ( sandford & allamandola 1993 ) , and a decrease in the gas - to - dust ratio is at odds with evidence that suggests an increase in high extinction regions ( e.g. bohlin et al . 1978 , de geus & burton 1991 , and ciolek & mouschovias 1996 ) . the density distribution in protostellar cores may also be studied through dust emission . if the emission is observed at long enough wavelength that the rayleigh - jeans approximation applies , then , in the optically thin regime , the intensity at impact parameter @xmath162 is proportional to the integral along the line - of - sight of the product of the density and the temperature . assuming that the intensity , temperature and density obey power law distributions in radius , then the power law index of the density distribution @xmath163 is simply given by @xmath164 , where @xmath113 is the observed power law index of the intensity profile , and @xmath165 the power law index of the temperature profile . the mass may also be derived from the radial intensity distribution , given an assumed mass opacity . unfortunately , the mass opacity of dust at millimeter and submillimeter wavelengths is uncertain , by a factor of five or more ( ossenkopf & henning 1994 ) . for b335 , dust emission at submillimeter and millimeter wavelengths has been studied recently by shirley et al . ( 2000 ) and motte & andr ( 2001 ) . shirley et al . ( 2000 ) constructed images at 850 @xmath20 m and 450 @xmath20 m at resolutions of @xmath166 and @xmath167 , respectively , using scuba at the jcmt . at both wavelengths , these observations are sensitive to emission out to an angular radius of roughly @xmath168 . they adopt a temperature profile with power law index @xmath169 for a region where the envelope is optically thin to the bulk of the infrared radiation ( e.g. emerson 1988 ; butner et al . they fit the circularly averaged maps at each wavelength , extending from inner cutoffs at angular radii of @xmath170 for the 850 @xmath20 m profile and @xmath171 for the 450 @xmath20 m profile . they derive power law indices of @xmath172 and @xmath173 from the 850 @xmath20 m and 450 @xmath20 m emission , respectively . they estimate a mass for b335 of 1.2 m@xmath14 using the total 850 @xmath20 m flux in a @xmath174 area and assuming a sphere of constant density . motte and andr ( 2001 ) constructed a 1.3 mm continuum image of b335 with @xmath175 resolution using the iram 30 m telescope , detecting emission out to an angular radius of @xmath174 . they adopt an isothermal temperature distribution and calculate a density power law index of @xmath176 . in this model , the total mass is @xmath177 . they note an asymmetry in the emission , with the emission being elongated perpendicular to the outflow axis . as in this extinction study , they perform fits to the off - outflow regions , and as a result increase the uncertainty in their fitted density index by 0.2 ( included in the quoted uncertainty ) . taking account the uncertainties , the density power law indices derived from the millimeter and submillimeter emission agree both with each other and with the results from dust extinction analysis ( fit a , @xmath178 ) . however , the extinction method provides a direct measurement of the distribution that suffers none of the added uncertainties associated with the temperature profile in the core . we present a near - infrared extinction study of the protostellar collapse candidate b335 using deep integrations from hst / nicmos supplemented by data from the w.m . keck observatory . in summary : 1 . we determine transformations between the nicmos filters f160w and f222 m and the standard h and k bands in the cit photometric system for a range of @xmath41 color from 0 to 3 . at the accuracy of our data , the derived transformations do not have significant slopes or color terms . 2 . the nicmos mosaic image of the central @xmath26 region of b335 shows a dramatic fall off in the number of stars detected towards the protostar location , where the extinction increases due to the central concentration of dense core material . the nicmos mosaic image contains a total of 119 stars detected in both filters , the innermost of which is at a radius of @xmath70 . the photometry shows a steep gradient in the @xmath41 colors toward the central protostar . in addition , there is a strong asymmetry in the colors superimposed on the overall gradient that coincides with the bipolar outflow . we compare a series of models of dense core stucture to the extinction data , including the previously proposed models of inside - out collapse derived from molecular line observations . the _ shape _ of the @xmath136 radial profile is well matched by the inside - out collapse model , in particular the density fall - off as @xmath129 for the envelope and a flattening within the purported infall zone . an unstable bonnor - ebert sphere with dimensionless outer radius @xmath0 provides an equally good description of the density profile , and is indistinguishable from the collapse model over the range in radius where there are stars to fit . the inside - out collapse model is not unique . the power law index for the density distribution derived from extinction is consistent with recent determinations from dust emission measurements at millimeter and submillimeter wavelengths . the bipolar outflow appears to be responsible for the dominant asymmetry in the extinction data . stars whose line of sight fall within @xmath137 of the outflow axis show lower extinction than those further away from the outflow . the effect of the outflow on the extinction data is reproduced very well by models that consider a hollowed out bipolar cone of constant opening angle . in the context of the inside - out collapse models , the observed asymmetry is too large to be explained by either rotation , or a weak magnetic field . 5 . the entire extinction dataset can be well fitted by a model that includes infall , bipolar outflow , and an additional @xmath140 dispersion that mimics residual turbulent motions . the best fit outflow semi - opening angle is @xmath3 and the best fit infall radius is @xmath1 . the fitted value of the infall radius is consistent with those derived from molecular line studies of b335 , and supports the inside - out collapse interpretation of the density structure . the fitted opening angle for the bipolar outflow is somewhat larger than has been observed in high velocity co emission , perhaps because the molecular line studies do not recover the full outflow width due to confusion with ambient emission at low velocities . the normalization of the observed color excess is a factor of @xmath179@xmath180 higher than the value predicted using the nominal 250 pc distance to b335 , and the standard gas - to - dust ratio and near - infrared reddening law . the discepancy can conceivably be explained by a combination of a closer distance and a lower gas - to - dust ratio . partial support for this work was provided by nasa through grant number hst - go-07843.01-a from the space telescope science institute , which is operated by aura , inc . , under nasa contract nas5 - 26555 we thank bob goodrich and randy campbell for invaluable assistance with the observations obtained at the w.m . keck observatory . dwah thanks brian mcleod for his help in running the nicreduce software , and alex wilson for his assistance with idl graphics . this appendix contains the tools and general recipe for constructing the simulated observations of b335 presented in sections [ sec : sims ] & [ sec : scaling ] . we construct a composite luminosity function for the background population by combining the data from fields away from the extinction of the b335 dense core . in deriving the luminosity function , the photometry for each field was cut at a very high completeness level , where the number of stars per unit magnitude range peaks . ( this is a very conservative estimate of limiting magnitude . ) to avoid any issues concerning how the stars should be binned , the cutoff was determined by finding where the gradient of the cumulative luminosity function was steepest . for the dataset from the fifth nicmos strip field , which includes 202 detected stars at h and 85 detected stars at k , the cut - off magnitudes are 20.7 at h , and 18.7 at k. for the combined dataset of the four keck off - fields , which includes 242 detected stars at h and 246 detected stars at k , the cut - off magnitudes were 18.8 for both filters . the empirical h - band luminosity function was calculated in three parts out to the nicmos high completeness limit ( h@xmath181 ) . for the bright end ( h@xmath182 ) , only stars from the nicmos fifth strip field are included , as stars brighter than this limit were saturated in the keck images . for the mid - range ( @xmath183 ) , stars from both datasets contribute to the function . for the faint end ( @xmath184 ) , above the keck high completeness limit , only stars in the nicmos fifth strip field were included . in each part , the number of stars was divided by the relevant observed area to build up the cumulative luminosity function . in the part where the nicmos and keck data overlap ( @xmath183 ) , the two datasets show good agreement . the k band luminosity function was constructed following the same procedure , and it is effectively the same as for the h band , offset by the mean color of @xmath185 magnitudes . the central regions of b335 were observed to a deeper limiting magnitude than the strip or the @xmath39 off fields , by about 4 magnitudes ( h band ) at the mosaic center , and about 3 magnitudes at the edge ( see figure [ fig : limplot ] ) . although the stars in these central regions of b335 are highly extincted , the empirical background luminosity function must be extrapolated in order to simulate the deepest observations . since the least extincted star in the nicmos mosaic has @xmath186 , corresponding to an extinction of @xmath187 , the luminosity function must be extrapolated by about 2 magnitudes at h in the simulations . the form of the h cumulative luminosity function suggests that a broken power law provides a good description . the function was fitted by minimizing the @xmath126 of the fit to the bright - end of the luminosity function . the resulting function is given by : @xmath188 simulated observations of b335 were constructed by superimposing a model cloud on a field of background stars with the appropriate characteristics . the column density was calculated at the position of each background star and converted into two observable quantities : ( 1 ) an extinction at h band , and ( 2 ) a color excess @xmath136 . these conversions are based on the gas - to - dust ratio , and reddening law . ( a brief discussion on the uncertainties of these quantities is given in section [ sec : method ] . ) 1 . generate a background star field by selecting randomly from the broken - power - law cumulative luminosity function in concert with the gaussian color distribution of the background population . the broken - power - law luminosity function is extended to h@xmath189 , just beyond the limiting magnitude at the center of the nicmos mosaic ( note that model background stars intrinsically fainter than h@xmath190 are not recovered in the simulated observations of the fitted models from table [ tab : fits ] the function is extended to h@xmath189 simply to ensure completeness ) . positions for each star are assigned randomly within a region @xmath191 , to encompass the area of the nicmos observations . 2 . select a model density distribution to describe b335 . 3 . calculate the column density for each star in the region . 4 . convert the column density to an extinction at h band , @xmath192 , and a color excess @xmath136 using an assumed gas - to - dust ratio , r , and a reddening law , @xmath193 . the `` standard '' prescription is ( bohlin et al . 1978 , rieke & lebofsky 1985 ) : @xmath194 { \mathrm{e}(\mathrm{h}-\mathrm{k})}\ : & = & a_{\mathrm{h } } \ : / \ : 2.78 \ .\end{aligned}\ ] ] ( in the simulations , the reddening law was allowed to vary up to 1525% steeper than the standard law in order to more exactly reproduce the observed starcounts . a fluctuation upward in the star counts at the one standard deviation level explains the inferred steeper reddening law . ) modify the magnitudes and colors of each background star appropriately . 6 . simulate the nicmos observations by applying the spatially dependent detection limits ( figure [ fig : limplot ] ) . llllll a & scaled power law : @xmath163 & @xmath195 & 146 & @xmath196 & 3.37 + b & scaled infall : @xmath197@xmath198 & @xmath199 & 146 & @xmath200 @xmath201 & 3.10 + c & scaled bonnor - ebert ( @xmath202 ) & @xmath203 & 146 & @xmath204 @xmath205 & 2.93 + d & scaled infall , hollow - cone outflow & @xmath206 & 209 & @xmath207 @xmath208 & 4.02 + e & as fit d with 20% dispersion & @xmath206 & 209 & @xmath209 @xmath210 & 1.04 +

we present a near - infrared extinction study of the dark globule b335 , a protostellar collapse candidate , using data from hst / nicmos and the w.m . keck observatory . these data allow a new quantitative test of the `` inside - out '' collapse model previously proposed to explain molecular line profiles observed toward this region . we find that the _ shape _ of the density profile is well matched by the collapse model , but that the _ amount _ of extinction corresponds to larger column densities than predicted . an unstable bonnor - ebert sphere with dimensionless outer radius @xmath0 provides an equally good description of the density profile , and is indistinguishable from the collapse model over the range in radius sampled by the extinction data . the bipolar outflow driven by the embedded young stellar object has an important effect on the extinction through the core , and modeling the outflow as a hollowed - out bipolar cone of constant opening angle provides a good match to the observations . the complete exinction map is well reproduced by a model that includes both infall and outflow , and an additional 20% dispersion that likely results from residual turbulent motions . this fitted model has an infall radius of @xmath1 ( @xmath2 pc for 250 pc distance ) , and an outflow cone semi - opening angle of @xmath3 . the fitted infall radius is consistent with those derived from molecular line observations and supports the inside - out collapse interpretation of the density structure . the fitted opening angle for the outflow is slightly larger than observed in high velocity co emission , perhaps because the full extent of the outflow cone in co becomes confused with ambient core emission at low velocities .

very high energy @xmath0 rays ( vhe ; energy @xmath1gev ) originating from extragalactic objects interact with photons of the extragalactic background light ( ebl ) leading to an exponential attenuation of the @xmath0-ray flux emitted by the source @xcite . direct detections of the ebl are difficult due to foreground emission @xcite and thus the exact level of the ebl photon density remains unknown . recent ebl models ( e.g. refs . @xcite ) predict densities close to lower limits deduced from galaxy number counts @xcite . in simple emission models of blazars - ray emitters , see e.g. http://tevcat.uchicago.edu/. ] , no spectral hardening is expected at vhe and consequently spectra corrected for ebl absorption should not show such features . nevertheless , evidence for such signatures has been found @xcite . an explanation might be the oscillations of @xmath0-rays into axion - like particles ( alps ) , spin-0 pseudo - nambu - goldstone bosons that arise in certain standard model extensions . these particles couple to photons in external magnetic fields ( see ref . @xcite for a review ) and could lead to a flux enhancement as alps do not interact with ebl photons . several turbulent magnetic - field environments have been studied in this respect including the intergalactic magnetic field @xcite , the agn host galaxy @xcite , intra - galaxy - cluster fields @xcite , magnetic fields in agn lobes @xcite , and in the milky way @xcite . the turbulent fields are commonly modelled with a simple domain - like structure : the path length is split up into @xmath2 domains of coherence length @xmath3 . while the field strength remains constant over all cells , the orientation of the field is assumed to change randomly from one cell to the next . the random nature of the field makes it necessary to calculate the conversion probability for a large number of random realisations . as @xmath2 grows , the photon - alp oscillations can often only be calculated numerically . here , we show analytically that in these simple models realisations exist for which the photon - alp conversion probability vanishes . the equations of motion of a monochromatic photon - alp beam composed of the two photon polarisation states @xmath4 and the alp field strength @xmath5 , @xmath6 , of energy @xmath7 propagating along the @xmath8 axis in a cold plasma with homogeneous magnetic field can be written as @xcite @xmath9 where the mixing is induced by off - diagonal elements of the mixing matrix @xmath10 . the resulting photon - alp oscillations are similar to neutrino oscillations and we denote the mixing angle by @xmath11 ( see , e.g. ref . @xcite for the full expressions for @xmath12 and @xmath11 ) . equation can be solved with the transfer matrix @xmath13 , so that @xmath14 , where @xmath15 denotes the angle between the transversal magnetic field and the photon polarisation state along @xmath16 @xcite . with the eigenvalues @xmath17 , @xmath18 , of the mixing matrix and introducing the notation @xmath19 and @xmath20 , the transfer matrix can be written as @xcite @xmath21 for @xmath2 consecutive domains with angle @xmath22 in each domain , it can be shown that the total transfer matrix is given as a product over all domains , @xmath23 present @xmath0-ray experiments can not measure the polarisation . therefore , one has to generalise the problem at hand to the density matrix formalism , where @xmath24 @xcite . the probability for an initially unpolarised photon beam , @xmath25 , to oscillate into an alp , @xmath26 is then given by @xcite @xmath27 the oscillation probability will take some value @xmath28 @xcite depending on the realisation of the angles @xmath29 and the mixing angle . interestingly , realisations exist for which @xmath30 even though @xmath31 . to show this we assume an even number of domains where @xmath32 in one half of the domains and @xmath33 in the other half ( where @xmath34 is a real non - zero number ) ordered randomly . a straightforward calculations shows that the commutator of the transfer matrices @xmath35 $ ] is an anti - symmetric matrix with entries @xmath36 and zero in all other entries . the matrix elements of the product @xmath37 that induce mixing ( i.e. , the @xmath38 , @xmath39 elements in the current basis ) are found to be equal to @xmath40 . above a critical energy @xmath41 the mixing becomes independent of energy . if in addition the mixing is strong so that @xmath42 , the commutator and the mixing inducing matrix elements vanish . with the commutator equal to zero we can now combine all pairs of @xmath43 and @xmath44 transfer matrices and see that the resulting product of all matrices given in eq . does not induce any photon - alp mixing . as an example , we show this behaviour in fig . [ fig : ex ] , in which we assume magnetic - field parameters found in galaxy clusters . the conversion probability is calculated numerically following eq . . above the critical energy the probability goes to zero , however , around the critical energy oscillations still occur . our findings still hold even if photon absorption is included as it is the case for conversions in the intergalactic magnetic field . however , as this magnetic field evolves with redshift , not all realisations lead to a conversion probability exactly equal to zero for all random permutations . as shown in the previous section , the photon - alp conversion probability can be exactly zero in special configurations of a turbulent magnetic field given that ( a ) it is modelled with a simple cell - like structure and ( b ) that the mixing occurs in the strong mixing regime , i.e. at energies @xmath45 and @xmath42 . oscillations around the critical energy still occur making spectral features at this energy a universal prediction of photon - alp oscillations . the absence of such signatures in @xmath0-ray spectra has already been used to constrain the photon - alp coupling @xcite . in more realistic models of the turbulent field ( that use , e.g. , a kolmogorov turbulence spectrum ) we do not have the freedom to choose the @xmath15 angles ( see , e.g. ref . @xcite ) and we can not easily construct a scenario with vanishing mixing as done here . utilizing such models , it can be shown that the future cherenkov telescope array will be sensitive to detect a boost in the photon flux for photon - alp couplings @xmath46 and alp masses @xmath47 @xcite , the same parameters that could explain evidence for a reduced opacity for vhe @xmath0-rays @xcite . mm is supported by a grant of the knut and alice wallenberg foundation , pi : jan conrad

axion - like particles ( alps ) are a common prediction of certain theories beyond the standard model and couple to photons in the presence of external magnetic fields . as a consequence , photon - alp conversions could lead to an enhancement of the flux of extragalactic @xmath0-ray sources that is otherwise attenuated due to the interactions with background radiation fields . the magnetic fields traversed by the @xmath0 rays are often turbulent and frequently modelled with a simple domain - like structure . given a maximum mixing between photons and alps , we show that in such models realisations of the fields exist for which the photon - alp oscillation probability vanishes . this behaviour does not occur in more sophisticated magnetic - field models .

determining the transport coefficients of many - particle systems is one of the fundamental problems of nonequilibrium statistical physics . this is also a notoriously difficult problem : it turns out that even in a simplified system where a single particle moves in a periodic array of scatterers , the drift and diffusion coefficients are highly irregular , apparently nowhere differentiable functions of control parameters @xcite ( an example is also shown in fig . [ fig1 ] below ) . closer inspection of these functions reveals that usually their `` irregularities '' are not random , but rather form patterns . this has led some researchers to the idea that the graphs of these functions are fractals with nontrivial , perhaps locally varying fractal dimension . this is an interesting concept , since fractal structures are quite commonly found in dynamical systems in various contexts @xcite . until very recently no reliable investigation of fractal properties of transport coefficients was possible because all general methods of calculating transport coefficients in dynamical systems e.g. the transition matrix technique combined with the escape rate formalism @xcite , the green - kubo formula @xcite , or the periodic - orbit formalism @xcite eventually lead to complicated and time - consuming numerical calculations . moreover , usually they are applicable only for some special values of the control parameters . for these reasons none of them could be used to collect a sufficiently large number ( counted at least in millions ) of very precise data required in fractal analysis . this situation changed when groeneveld and klages @xcite gave exact formulas for the transport coefficients in a simple one - dimensional dynamical system introduced by grossmann and fujisaka @xcite . the system investigated by groeneveld and klages can be considered as a model of a particle moving in a one - dimensional array of scatterers . the role of the equation of motion is played by a one - dimensional map @xmath0 @xmath1 where @xmath2 is a discrete - time variable and the map @xmath3 is given by a simple linear function @xmath4 on the fundamental interval @xmath5 , with @xmath6 being the control parameters representing the slope and the bias of the map , respectively ; the map is then continued periodically onto the real line by a lift of degree one , i.e. by requiring that @xmath7 for @xmath8 the lyapunov exponent of this system is positive , and so the dynamics defined by @xmath9 is chaotic . if we choose an arbitrary number @xmath10 as the starting point , the resulting sequence @xmath11 will almost always look `` random '' ( the set of points @xmath10 generating a regular , periodic or quasi - periodic sequence @xmath11 is of lebesgue measure 0 ) . for suitably chosen values of @xmath12 and @xmath13 the deterministic dynamics defined by the map @xmath9 is equivalent to a markov stochastic process of random - walk type : each deterministic trajectory @xmath11 is equivalent to some particular realization of the corresponding random - walk process , and taking the average over all initial states is equivalent to calculating averages over the corresponding gibbs ensamble . actually any random - walk markov process in a one - dimensional periodic system with fixed transition rates can be translated into the language of simple piecewise linear deterministic maps @xcite . from this point of view there is no surprise that the process defined by @xmath9 ( or similar maps ) is called `` deterministic diffusion '' and that the two basic transport coefficients , the drift velocity @xmath14 and diffusion constant @xmath15 , can be defined as @xmath16 where @xmath17 denotes the average over the uniform ensamble of initial values @xmath10 . the graph of the diffusion coefficient @xmath15 for the map @xmath9 as a function of the slope @xmath12 for the bias @xmath18 is shown in fig . the fractal properties of this highly irregular graph ( as well as that of the drift velocity @xmath14 ) were recently studied by klages and klau@xcite . using two numerical methods : the box counting and the autocorrelation function methods , they found that the local fractal dimensions of graphs of @xmath14 and @xmath15 are well - defined , but highly irregular functions of @xmath12 and @xmath13 . in other words they found that the graph shown in fig . [ fig1 ] can not be described with a single fractal dimension , but rather by a set of quickly varying local fractal dimensions . taking this into account they suggested that the local fractal dimensions of graphs of @xmath14 and @xmath15 as functions of the slope @xmath12 with the bias @xmath13 fixed are fractal themselves . this , in turn , leads to the concept of a `` fractal fractal dimension of deterministic transport coefficients '' @xcite . the diffusion coefficient @xmath15 for the map @xmath9 as a function of the slope @xmath12 for the bias @xmath18 . the inset depicts a blow - up of a region @xmath19 ( pointed at by the arrow).,scaledwidth=50.0% ] klages and klau found the local fractal dimension @xmath20 of @xmath21 and @xmath22 to be very close to 1 . the autocorrelation function method gave @xmath23 , and the box - counting method gave even lower values @xmath24 . all these values are very close to 1 , i.e. the fractal dimension of a regular , `` smooth '' curve . this suggests that perhaps @xmath25 and that the results obtained in the above - mentioned computer simulations reflected extremely slow convergence of @xmath20 to its true asymptotic limit of infinitesimally small `` boxes '' . such slow convergence is often caused by logarithmic corrections . the main purpose of my paper is thus to try and find such tiny corrections . to accomplish this task i will use a different method of calculating the local fractal dimension of a curve the so called oscillation method @xcite and i will carry out all calculations with a help of a special high - accuracy numerical library , which will enable me to consider exceptionally small `` boxes '' . for sake of simplicity i shall restrict the present study to the simplest case of zero bias ( @xmath18 ) where , due to symmetry , the drift velocity @xmath26 . the structure of the paper is as follows . section [ sectionmethod ] briefly describes the method i have used to determine the local fractal dimension . it describes all steps necessary to calculate the diffusion coefficient for the map @xmath9 , a method of calculating the local fractal dimension of a continuous curve ( `` the oscillation method '' ) , mathematical formulation of the main conjecture about the logarithmic convergence of the fractal dimension to its limiting value , and the numerical aspects of the algorithms used . section [ sectionresults ] presents the main results . finally , section [ sectionconclusions ] is devoted to discussion of results . the explicit formulas @xcite for the transport coefficients in the model depend on some auxiliary variables . for each @xmath27 , and @xmath28 we define two infinite sequences , @xmath29 , @xmath30 , and @xmath31 , @xmath32 , consisting of real and integer numbers , respectively . their values are uniquely determined by demanding that @xmath33 and that for each @xmath34 @xmath35 with additional conditions @xmath36 and @xmath37 , where @xmath38 $ ] and @xmath39 . next we define `` n - numbers '' : @xmath40 @xmath41 r^l,\ ] ] where @xmath42 . the basic transport coefficients , @xmath14 and @xmath15 , can be now expressed ( @xcite , cf @xcite ) as @xmath43 the fractal dimension @xmath20 of a continuous function @xmath44 defined on an interval @xmath45 $ ] can be evaluated through analysis of its hlder exponents @xcite and @xmath46-oscillations @xcite . the @xmath46-oscillation of @xmath44 at @xmath47 is defined as @xmath48 if there exist constants @xmath49 and @xmath50 such that for all @xmath46 @xmath51 then @xmath44 is called a holderian of exponent @xmath52 at @xmath53 and its fractal dimension @xmath20 at @xmath53 is related to @xmath52 through @xmath54 similarly , if there exist constants @xmath49 and @xmath50 such that for all @xmath46 @xmath55 then @xmath44 is called an anti - holderian of exponent @xmath52 at @xmath53 and @xmath56 we are now ready to formulate our main conjecture : for the map @xmath9 with @xmath18 @xmath57^{\gamma(a)},\quad \mathrm{as~ } \tau \to 0,\ ] ] where the prefactor @xmath58 and the exponent @xmath59 . owing to ( [ def - osc ] ) ( [ deltage ] ) this implies that the fractal dimension of the graph of @xmath15 as a function of the control parameter @xmath12 is equal 1 , but the convergence to this limiting value is logarithmically slow , with the exponent @xmath60 controlling the convergence rate . i checked conjecture ( [ conjecture ] ) using equations ( [ 2reccursion ] ) ( [ jd ] ) . because the consecutive terms of the sequences @xmath61 in ( [ nr ] ) can be generated extremely efficiently ( it took less than a second to generate 20 000 data points used to draw figure [ fig1 ] ) , i decided to employ gmp ( gnu multiple precision arithmetic library , version 4.1.2 ) a library for arbitrary precision arithmetic @xcite . although any calculations performed by such a library must be several orders of magnitude slower that those performed directly , thanks to the gmp i was able to study numerically the limit of @xmath62 for @xmath46 ranging from @xmath63 down to at least @xmath64 ( the accuracy of typical computers currently available , without using such special - purpose libraries , is limited typically to about @xmath65 ) . such fine resolution of my calculations will turn out crucial for determining logarithmic corrections predicted by formula ( [ conjecture ] ) . in my present study i decided to concentrate on verifying conjecture ( [ conjecture ] ) for the symmetrical case of vanishing bias ( @xmath66 ) and for several selected values of the slope @xmath12 . i decided to examine especially carefully the slopes corresponding to finite markov partitions . these values are particularly interesting from the theoretical point of view @xcite ; for example , even though the diffusion coefficient @xmath15 is always continuous in @xmath12 @xcite , the numerator and denominator in ( [ nkl : eps ] ) can be discontinuous for such ( and only such ) slopes . one of the reasons why i chose to set @xmath18 is that in this case the slopes corresponding to markov partitions can be found easily on a computer : they are algebraic numbers generating periodic sequences @xmath67 and @xmath68 . such values of @xmath12 will be henceforth referred to as `` markov slopes '' . figure [ fig2 ] presents results obtained for two markov slopes generating strictly periodic sequences ( i.e. , @xmath69 for a period @xmath70 and all @xmath71 ) . the value used in fig . [ fig2]a is @xmath72 . this slope generates simple sequences @xmath73 and @xmath74 of period 1 . the oscillations of the diffusion coefficient on intervals @xmath75 , rescaled by @xmath46 , obtained for @xmath76 , are represented in this plot by circles , while the solid line represents a quadratic fit of form @xmath77 with @xmath78 , @xmath79 , and @xmath80 . the fit is excellent , which suggests that for this slope the exponent @xmath81 , as defined in eq . ( [ conjecture ] ) , is 2 . i have obtained similarly good quadratic fits for other integer odd slopes , which also generate simple sequences of period 1 @xcite . oscillations of the diffusion coefficient rescaled in accordance with eq . ( [ conjecture ] ) calculated on intervals @xmath82 for @xmath83 and for ( a ) @xmath84 and ( b ) @xmath85 . the circles represent results calculated from conjecture ( [ conjecture ] ) , and the solid lines are quadratic fits . the inset in graph ( b ) shows the blow - up of results obtained for @xmath86 . ] a question arises whether the length of the period has any influence on the limiting value of @xmath87 . figure [ fig2]b shows the results obtained for the largest root of the polynomial @xmath88 , i.e. for @xmath89 . for this slope the period @xmath90 . just as in the previous example , @xmath87 can be approximated by a quadratic , although the fit is not as excellent as for @xmath72 . the inset in this figure depicts the blow - up of the data obtained for @xmath91 . it shows that for @xmath92 the value of @xmath87 fluctuates about a constant value , which might suggest that the exponent @xmath81 vanishes . this example demonstrates that the `` resolution '' of calculations used in ref.@xcite , e.i . @xmath93 , is insufficient to find the true asymptotic fractal properties of the graph of the diffusion coefficient as a function of the slope @xmath12 . the same as in fig . ( a ) @xmath94 ( circles ) , @xmath95 ( squares ) , @xmath96 ( pluses ) with linear fits ( solid lines ) ; ( b ) @xmath97 ( circles ) , @xmath98 ( squares ) with quadratic fits ( solid lines ) . see text for details . ] all other markov slopes generate sequences @xmath99 and @xmath100 with the period starting at a term @xmath101 . there are two cases : either all terms of @xmath102 are different from all the terms of @xmath103 or the two sequences have the same period sequence . let s start from the latter case . three examples of results obtained for markov slopes of this type are shown in figure [ fig3]a . the data depicted in this graph were calculated for @xmath104 ( circles ) , for the largest root of @xmath105 , i.e. for @xmath106 ( squares ) , and for the largest root of @xmath107 , i.e. for @xmath96 ( pluses ) . for the first two of the above slopes all the terms of the sequences @xmath100 eventually vanish . in the case of @xmath104 this happens for @xmath108 , and for @xmath109 the terms of @xmath110 vanish for @xmath111 . as for the third example , @xmath112 , the period starts at @xmath113 and has the length 6 . specifically , the first five terms of @xmath102 are ( to 3 s.f . ) @xmath114 , @xmath115 , @xmath116 , @xmath117 , @xmath118 , and for @xmath119 the terms satisfy @xmath120 . as we can see , in all these cases @xmath87 diverges linearly with @xmath121 , suggesting that conjecture ( [ conjecture ] ) is satisfied with the exponent @xmath122 . the results obtained for the last category of markov slopes , i.e. slopes generating two disjoint periodic sequences @xmath102 and @xmath103 , are shown in fig . the two data sets were collected for the largest root of @xmath123 , i.e. for @xmath124 ( circles ) and for the largest root of @xmath125 , i.e. for @xmath126 ( squares ) . the first of these slopes , @xmath127 , generates a sequence with a period of length 3 starting at @xmath128 , while @xmath126 generates a sequence of period length 4 starting at @xmath129 . as we can see , for these markov slopes @xmath87 can be very well approximated by a quadratic function of @xmath121 . this suggests that in this case conjecture ( [ conjecture ] ) is satisfied with the exponent @xmath130 . figure [ fig4 ] presents results for two slopes @xmath12 which do not correspond to a markov partition . these are @xmath131 ( left panel ) and @xmath132 ( right panel ) . the results turned out to be `` noisy '' and i had to increase the resolution down to @xmath133 . as can be seen , in both cases @xmath87 has a clear trend linear in @xmath121 . however , this limiting behaviour is disturbed by very large fluctuations . actually , for @xmath134 these fluctuations are so large that my numerical results can not rule out the possibility that in this case the limit @xmath135 does not exist . the same as in fig . [ fig2 ] for two slopes @xmath12 that do not correspond to finite markov partitions : ( a ) @xmath131 and ( b ) @xmath136 . ] in my study i proposed and verified numerically a hypothesis that the local fractal dimension of the graph of the diffusion coefficient of the piecewise linear map @xmath9 as a function of the slope @xmath12 is @xmath25 and that the convergence to this limit is slowed down by logarithmic corrections described by eq . ( [ conjecture ] ) . this contradicts the earlier findings of klages and klau@xcite that this graph is a fractal with a locally varying fractal dimension @xmath137 which , when plotted as a function of the slope , forms a fractal itself . i found that the exponent @xmath81 , which controls the logarithmic correction , is actually a function of the slope @xmath12 . interestingly , @xmath81 appears to be a discontinuous function that can take only one of two values : 1 or 2 . the value of 1 corresponds to markov slopes that generate two disjoint sequences @xmath102 and @xmath103 , and the value of @xmath138 corresponds to markov slopes that generate periodic sequences @xmath102 and @xmath103 with the same period terms . the case of slopes that do not generate markov partitions is not clear apparently @xmath122 , but this statement can not be verified numerically because of very large fluctuations making the convergence extremely slow . note that my findings imply that both of the sets @xmath139 and @xmath140 are dense , and hence that @xmath141 is nowhere continuous . any numerical study is naturally restricted to investigation of several particular cases . it can not be ruled out that i have overlooked some categories of slopes @xmath12 for which the local fractal dimension of the graph of the diffusion coefficient behaves in a way not predicted in this study . numerical investigation of the logarithmic correction to the asymptotic limit is here a particularly delicate problem , as apparently @xmath81 is a nowhere continuous function of the slope . only through an analytical approach could this problem be solved conclusively . such a study will be published elsewhere . [ sectionconclusions ] i thank r. klages for many inspiring discussions . support from the polish kbn grant nr 2 p03b 030 23 is gratefully acknowledged .

using a numerical library for arbitrary precision arithmetic i study the irregular dependence of the diffusion coefficient on the slope of a piecewise linear map defining a dynamical system . i find that the graph of the diffusion coefficient as a function of the slope has the fractal dimension 1 , but the convergence to this limit is slowed down by logarithmic corrections . the exponent controlling this correction depends on the slope and is either 1 or 2 depending on existence and properties of a markov partition .

a broad enhancement @xmath0 with @xmath6 was observed by bes collaboration in @xmath7 invariant mass spectrum in the @xmath8 channel @xcite . its resonance parameters are @xmath9 mev . the branching ratio is @xmath10\cdot b[x(1576)\to k^{+}k^{-}]=(8.5\pm 0.6^{+2.7}_{-3.6})\times 10^{-4}$ ] . its extremely large width around 800 mev motivated theoretical explanations such as a @xmath11 molecular state @xcite , tetraquark @xcite , diquark - antidiquark bound state @xcite , a composition of @xmath12 and @xmath13 @xcite . since there are two broad overlapping resonances @xmath12 and @xmath13 with the same quantum number around 1600 mev , we investigated whether such a broad signal could be produced by the final state interaction ( fsi ) @xcite effect in our previous work @xcite . we noticed that the interference effect of @xmath2 could produce an enhancement around @xmath14 mev with the opening of the @xmath15 channel , similar to the cusp effect discussed in ref . however , the branching ratio @xmath16\cdot b[\rho(1450,1700)\to k^{+}k^{-}]$ ] from the fsi effect was far less than the experimental data . it s important to point out that we considered only the contribution of the absorptive part in ref . @xcite . recently meng and chao explored the possible assignment of @xmath17 as the @xmath18 candidate @xcite . they found the dispersive part of the fsi amplitude contributes more importantly to the hidden charm decay width of @xmath17 than the imaginary part derived in ref . @xcite because the intermediate states @xmath19 and @xmath20 lie very close to @xmath17 . motivated by the above observation , we investigate the potential role of the dispersive contribution of fsi since the @xmath15 intermediate states are rather close to @xmath2 . this paper is organized as follows . the formulation of @xmath21 by exchanging @xmath22 is presented in section [ sec2 ] and our numerical result and discussion in section [ sec3 ] . as shown in fig . [ kk ] , we focus on the fsi of the @xmath15 intermediate states through the exchange of the @xmath22 meson : @xmath21 . [ cols="^,^ " , ] c in this short note , we revisit the final state interaction of the @xmath5 decay . different from our former work @xcite , we consider the additional contribution from the dispersive part . our result shows that there exists an enhancement around 1580 mev as shown in fig . [ diagrams ] . such an enhancement occurs with the opening of the @xmath15 channel . the decay width of @xmath23 from the fsi effect is a few mev only . if the width of @xmath2 is 300 mev , the branching ratio of @xmath24 is about @xmath25 . with @xmath26 $ ] roughly around @xmath27 @xcite , @xmath28\cdot b[\rho(1450,1700)\to ab\to k^{+ } { k}^{-}]$ ] is about @xmath29 . the dispersive contribution enhances the branching ratio of @xmath30 by two orders than that in ref . however , such a ratio is still far less than experimental value @xmath31\cdot b[x(1576)\to k^{+}{k}^{-}]=(8.5\pm 0.6^{+2.7}_{-3.6})\times 10^{-4}$ ] , although the @xmath32 spectrum from the fsi effect of @xmath2 decays mimics the observed broad spectrum from bes s measurement . throughout our calculation , we ignored the direct coupling between @xmath33 and @xmath34 . recently , li argued that @xmath33 can have strong coupling with @xmath7 at the tree level @xcite . adding this contribution certainly increases the branching ratio . however , the experimental upper limit of @xmath35 $ ] is @xmath36 and @xmath34 is not one of the dominant decay modes of @xmath13 @xcite . clearly , future besiii high - statistics data around 1.6 gev in the @xmath34 channel will be very helpful in the clarification of @xmath0 . we also calculate the decay amplitude of @xmath37 using the same technique . there exists one similar enhancement around 1580 mev , which is shown in fig . [ diagrams-1 ] . this enhancement will be useful to test if x(1576 ) arises from the fsi effect . we thank prof . b.s . zou for the useful suggestion . this project was supported by the national natural science foundation of china under grants 10421503 and 10625521 , and the china postdoctoral science foundation ( 20060400376 ) . n. isgur , k. maltman , j. weinstein and t. barnes , phys . lett . * 64 * , 161 ( 1990 ) ; m.p . locher , v.e . markusin and h.q . zheng , report no . psi - pr-96 - 13 ( unpublished ) ; h. lipkin , nucl . phys . * b 244 * , 147 ( 1984 ) ; h. lipkin , phys . lett . * b 179 * , 278 ( 1986 ) ; h. lipkin , nucl . * b 291 * , 720 ( 1987 ) .

we study whether the broad enhancement @xmath0 arises from the final state interaction ( fsi ) of @xmath1 decays . we consider both the absorptive and dispersive contribution of the above amplitudes since the intermediate states are very close to @xmath2 . the same mechanism leads to a similar enhancement around 1580 mev in the @xmath3 spectrum in the @xmath4 channel , which can be used to test whether @xmath0 can be ascribed to the fsi effect of @xmath5 .

the introduction of a three - nucleon force ( 3nf ) arising from the exchange of two pions among three nucleons ( 2@xmath0e ) as shown in fig . [ fig:3nf - graph ] ( a ) into nuclear hamiltonian is known to get rid of discrepancies between experimental data and theoretical calculations with realistic two - nucleon forces ( 2nfs ) for the three - nucleon ( 3n ) binding energies and nucleon - deuteron ( nd ) differential cross sections . on the other hand , the 2@xmath0e 3nf is unsuccessful in explaining some nd polarization observables , _ e.g. _ , too small effects to vector analyzing powers ( vaps ) and undesirable contributions to tensor analyzing powers ( taps ) in low - energy nd elastic scattering ( see fig . [ fig : t21-br-3 - 28mev ] ) . in refs . @xcite , we have pointed out that tensor components in the 2@xmath0e 3nf should be responsible to the problem in taps . in this paper , we examine a 3nf due to the exchange of a pion and a scalar - isoscalar object ( @xmath0-s ) shown in fig . [ fig:3nf - graph ] ( b ) as a possible source of tensor interactions that have different characteristics from those of the 2@xmath0e 3nf . after giving a general form of the @xmath0-s 3nf in sec . [ sec : ps-3nf ] , numerical results for the 3n binding energy and nd scattering observables will be presented in sec . [ sec : results ] . summary is given in sec . [ sec : summary ] . of @xmath1 elastic scattering at @xmath2 mev and @xmath3 mev . the data are taken from refs . @xcite for 3 mev and from ref . @xcite for 28 mev . [ fig : t21-br-3 - 28mev ] , title="fig : " ] of @xmath1 elastic scattering at @xmath2 mev and @xmath3 mev . the data are taken from refs . @xcite for 3 mev and from ref . @xcite for 28 mev . [ fig : t21-br-3 - 28mev ] , title="fig : " ] of @xmath1 elastic scattering at @xmath2 mev and @xmath3 mev . the data are taken from refs . @xcite for 3 mev and from ref . @xcite for 28 mev . [ fig : t21-br-3 - 28mev ] ] we will consider the following models for the @xmath0-s 3nfs : @xmath0-@xmath4 exchange with the excitation of the roper resonance @xmath5 ( @xmath6-@xmath7 ) @xcite ; @xmath0-@xmath4 exchange corresponding to the nucleon born diagrams ( so called pair or z diagrams ) for the pv ( @xmath6-@xmath8 ) or the ps ( @xmath6-@xmath9 ) @xmath10 coupling @xcite ; @xmath0-``effective scalar field '' exchange by a linear model @xcite or by a nonlinear model @xcite ; @xmath0-``effective 2@xmath0 '' exchange ( @xmath6-@xmath11 ) @xcite . a coordinate space representation of these potentials corresponding to diagram fig . [ fig:3nf - graph ] ( b ) is : @xmath12 y_\pi(r_{12 } ) y_\sigma(r_{32 } ) , \label{eq : wps}\ ] ] where @xmath13 with @xmath14 being a position vector of nucleon @xmath15 . when a dipole @xmath16 form factor with a cutoff mass @xmath17 is used , the function @xmath18 becomes @xmath19 . expressions of @xmath20 and @xmath21 for the @xmath0-s 3nfs @xcite are shown in table [ tab : pot_strength ] . .[tab : pot_strength ] the potential strengths @xmath20 and @xmath21 for the @xmath0-s 3nf models and and effects of the 3nfs on the @xmath22h energy : @xmath23 . [ cols="<,^,^,^,^,^",options="header " , ] in the present calculations , we use the following parameters : @xmath24 ; another coupling constants from table iii of ref . @xcite ; @xmath25 mev ; a set of @xmath26 for the @xmath27 ) 3nf @xcite ; @xmath28 mev and @xmath29 mev . first , we note that the argonne @xmath30 ( av18 ) 2nf @xcite underbinds the triton ( @xmath22h ) by 0.85 mev , and the brazil 2@xmath0e ( br@xmath31 ) 3nf @xcite gives an additional attraction of -1.75 mev . thus a repulsive effect is expected to the @xmath0-s 3nf to complete the nuclear hamiltonian . calculated values of the @xmath22h energy for the av18 plus each of the @xmath0-s 3nf models are presented in table [ tab : pot_strength ] as differences from the av18 calculation . this shows that a @xmath0-s 3nf with ( positive ) @xmath20-term produces an attractive contribution to the @xmath22h energy , and that with ( negative ) @xmath21-term a repulsive contribution , which is consistent with the results given in ref . the @xmath6 - 2@xmath32 3nf consists of a negative @xmath20-term and a negative @xmath21-term , and produces a repulsive effect mostly due to the @xmath20-term . in order to investigate effects of each term in the @xmath0-s 3nf on nd scattering observables , we pick up the following two @xmath0-s 3nf models to reproduce the @xmath22h energy together with the av18 2nf , the br@xmath33 3nf , and a phenomenological spin - orbit type 3nf ( so ) @xcite : @xmath34 the @xmath6-@xmath8 3nf as a representative to @xmath21 term ; @xmath34 the @xmath6 - 2@xmath32 3nf as a representative to @xmath20 term although it includes a small effect from @xmath21 term . ( in this case , we take @xmath35 mev . ) the inclusion of the so 3nf is effective to reproduce the vaps , but gives only minor effects on the @xmath22h energy and the taps at low energies . numerical results of faddeev calculations @xcite for the tensor analyzing power @xmath36 at @xmath37 mev and @xmath38 mev are presented in fig . [ fig : t21-ps-3 - 28mev ] , where we plot the experimental data and the results with the 3nfs divided by the av18 calculations . for @xmath36 at 3 mev , both @xmath0-s 3nfs equally tend to cancel the effect due to the 2@xmath0e 3nf opposite to the data , but still leave an amount of discrepancy . at 28 mev , the 3nfs contribute to @xmath36 in different ways depending on scattering angles . at scattering angles of @xmath39 to @xmath40 , the calculation with the br@xmath31 3nf and the one with the br@xmath31 + @xmath41 3nfs look consistent with the data . on the other hand , at scattering angles of @xmath42 to @xmath43 , the calculation with the br@xmath31 + @xmath6-@xmath8 3nfs as well as the one with only 2nf look consistent with the data . in fig . [ fig : del - sig - tl - ps ] ( a ) , we display results of the transversal @xmath44 and the longitudinal @xmath45 asymmetries of the spin dependent total cross sections in @xmath46 scattering comparing with recent experimental data of @xmath45 @xcite . in ref . @xcite , we showed that effects of tensor interactions are prominent in the difference of @xmath47 . as expected , differences of tendency in tensor components of the 3nfs are significantly observed in fig . [ fig : del - sig - tl - ps ] ( b ) . and @xmath45 ( a ) , and their difference ( b ) . the data of @xmath45 are taken from ref . @xcite . [ fig : del - sig - tl - ps ] ] and @xmath45 ( a ) , and their difference ( b ) . the data of @xmath45 are taken from ref . [ fig : del - sig - tl - ps ] ] we have examined three - nucleon forces arising from the exchange of @xmath0 and scalar - isoscalar object among three nucleons : @xmath0-@xmath4 exchange via z - diagram ( @xmath21-term in eq . ( [ eq : wps ] ) ) ; @xmath0-effective 2@xmath0 exchange ( @xmath20-term , small @xmath21-term ) . both models produce repulsive effects on the @xmath22h energy to compensate strong attraction caused by the 2@xmath0e 3nf . each 3nf affects polarization observables of nucleon - deuteron scattering , @xmath36 and @xmath48 , in different ways . this shows that further measurements of these observables at some energies provide additional information on three - nucleon forces , which should be included in the nuclear hamiltonian in addition to the 2@xmath0e 3nf . s. ishikawa , m. tanifuji , and y. iseri , proc . of the seventeenth international iupap conference on few - body problems in physics . durham , north carolina , usa , june 5 - 10 , 2003 , edited by w. glckle , w. tornow , ( elsevier , 2004 ) s61 .

effects of three - nucleon forces arising from the exchange of a pion and a scalar - isoscalar object among three nucleons on nucleon - deuteron scattering observables are studied .

the primary motivation for supersymmetry ( susy ) is to stabilize the smallness of the electroweak scale against radiative corrections @xcite , which can be as large as the planck scale if the higgs bosons are truly elementary . once the electroweak scale is set in the tree - level lagrangian , it only receives logarithmic radiative corrections , and hence its order of magnitude is not changed . moreover , the electroweak symmetry remains unbroken in the minimal supersymmetric standard model ( mssm ) in the absence of explicit susy - breaking parameters . therefore , one can view the electroweak symmetry breaking as being triggered by the soft susy breaking . indeed , the soft susy - breaking mass - squared of the higgs boson can be driven negative due to the top quark loop @xcite while all the other scalar bosons still have positive mass - squared . in this sense , there is nothing special about the higgs boson . it is just one of many scalar bosons , which happens to acquire a negative mass - squared due to the top quark loop . this idea eliminates one of the least appealing features of the standard model . however , there are at least two open questions . first , susy by itself does not explain why the electroweak scale is small to begin with . therefore , susy makes the smallness of the electroweak scale `` technically natural , '' but not truly natural . second , the mssm contains one dimensionful parameter ( the @xmath0-parameter ) , allowed by susy , in the superpotetial . the natural values of @xmath0 are either the planck mass ( the only natural dimensionful parameter available ) or zero , but recent experimental constraints imposed by lep2 imply that a nonzero @xmath3 gev is required @xcite . susy , fortunately , can potentially explain the smallness of the electroweak scale if it is broken dynamically @xcite . the perturbative non - renormalization theorem forbids the generation of a mass scale in the superpotential if it is absent at the tree - level . however , non - perturbative effects can violate the non - renormalization theorem , and a mass scale can be generated by a dimensional transmutation : @xmath4 , if an asymptotically free gauge theory is responsible for susy breaking . there has been major progress in building models of dynamical susy breaking @xcite , which became possible with the detailed understanding of the non - perturbative dynamics of susy gauge theories @xcite . furthermore , the so - called gauge mediation of susy breaking ( gmsb ) @xcite can generate soft susy - breaking parameters in the susy standard model in a phenomenologically desired form . therefore , there is hope of understanding the smallness of the electroweak scale in a truly natural manner . however , the other question remains largely unanswered : how can the dimensionful parameters in the superpotential naturally be of the order of the susy - breaking parameters ? there have been extensive discussions on this subject in the literature which we briefly summarize in section 2 . unfortunately , many of the proposed mechanisms rely on either small parameters , accidental cancellations , or the absence of interactions allowed by symmetries . we find the current situation to be rather unsatisfactory . a natural direction to follow is to start with a superpotential which does not contain a dimensionful parameter and hope that the electroweak scale is generated solely due to the soft susy - breaking parameters . the simplest model which can potentially work along this line is the next - to - minimal supersymmetric standard model ( nmssm ) @xcite , which replaces the @xmath0-parameter by the vacuum expectation value of an electroweak singlet superfield . we revisit this possibility with detailed quantitative studies in this paper . unfortunately , our conclusion is negative . the nmssm by itself does not produce a phenomenologically viable electroweak symmetry breaking even if we vary the messenger scale . the major experimental constraints include higgs boson and slepton searches . certain simple modifications can evade phenomenological constraints , but require a cancellation among parameters accurate to a few percent . we present all of these points quantitatively in this paper , and hope that our results prompt further investigations in understanding the origin of the @xmath0-parameter in models with the gmsb . the paper is organized as follows . in the next section , we review the situation of the @xmath0-problem in models with the gmsb , and discuss various proposals to explain the origin of the @xmath0-parameter . we , however , find none of them entirely satisfactory . even if we accept one of the proposed models , it is still necessary to check whether the generated @xmath0-parameter is phenomenologically allowed . we address this question in section 3 , and find that the currently available experimental lower bounds on superparticle masses already require a cancellation of order @xmath1% between the @xmath0-parameter and soft susy - breaking parameters to reproduce the observed @xmath5 . then , as the major part of our study , we present the quantitative results of the electroweak symmetry breaking in the nmssm with the gmsb in section 4 and find that there is no phenomenologically viable parameter set even if we vary the messenger scale from @xmath6 to @xmath7 gev . we study various simple modifications of the nmssm in section 5 , and find that they either do not break electroweak symmetry in a phenomenologically viable manner or require a cancellation among parameters of order @xmath8 . we finally conclude in section 6 . in this section , we review the @xmath0-problem in the supersymmetric standard model in general , and also various attempts to solve it in the context of the gmsb . the parameter @xmath0 is the only dimensionful quantity present in the superpotential of the mssm @xmath9 here , @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath14 are the matter chiral superfields with the obvious notation , and @xmath15 , @xmath16 the higgs doublets . note that @xmath0 is part of the supersymmetric lagrangian , and hence its origin is , naively , unrelated to the origin of the soft susy - breaking terms @xmath17 phenomenology , on the other hand , dictates that the values of both @xmath0 and the soft susy - breaking masses should be around the weak scale ( 100 gev ) , if susy is to be responsible for stabilizing the higgs mass . therefore , the important question is how the mechanism of susy breaking can induce a @xmath0-parameter naturally , at the same order of magnitude as the other soft susy - breaking parameters in the lagrangian . one popular scenario of susy breaking is the so - called `` hidden sector '' susy breaking in supergravity ( sugra ) @xcite . in hidden sector models , susy is broken in the hidden sector by some mechanism , such as the polonyi model @xcite , gaugino condensation @xcite , or the orafeartaigh model @xcite , and the effects of susy breaking are mediated to the fields in the supersymmetric standard model only by interactions suppressed by the planck scale . it therefore requires susy breaking at a scale @xmath18 gev if the soft susy - breaking masses are generated as @xmath19 . this class of models is able to generate the appropriate soft susy - breaking masses and @xmath0-parameter given that the @xmath0-term is forbidden in the supersymmetric limit by appropriate symmetries , and arises due to susy breaking ( see , for example , the giudice masiero mechanism @xcite ) . hidden sector models have , on the other hand , to face serious bounds imposed by flavor - changing neutral currents ( fcnc ) @xcite . low - energy constraints such as the smallness of @xmath20@xmath21 mixing require the matrices @xmath22 , @xmath23 to have eigenvalues degenerate to a few percent , or their eigenvectors to be strongly `` aligned '' with the eigenvectors of the yukawa matrices @xmath24 ( the same is true for @xmath25 ) . within the sugra framework alone , there is no natural mechanism to guarantee the degeneracy or the alignment @xcite . in this case , flavor symmetries are probably necessary to ensure either degeneracy @xcite or alignment @xcite and suppress fcnc , and some of the models presented are also capable of generating the @xmath0-term through flavor symmetry breaking @xcite . there is also the possibility that string theory generates degenerate squark masses if , for instance , the dilaton field provides the dominant contribution to the soft susy - breaking masses @xcite . the gauge mediation of supersymmetry breaking is an alternative mechanism which can naturally ensure the degeneracy of squarks masses and therefore suppress the dangerous fcnc effects . susy is somehow broken ( hopefully dynamically via dimensional transmutation to generate a large hierarchy ) , and susy - breaking effects are mediated to the fields in the supersymmetric standard model by the standard model gauge interactions . mediating susy breaking via gauge interactions is not a novel idea @xcite . it allows for susy breaking at a lower scale ( when compared to sugra inspired models ) and , because all susy - breaking effects are transmitted by flavor blind interactions ( the standard model gauge interactions ) , squarks of different families have the same mass . this scheme has attracted a lot of interest after the pioneering works by the authors of references @xcite , which showed that one can successfully mediate the susy - breaking effects via gauge interactions with the help of a so - called `` messenger sector . '' their scheme can easily incorporate dynamical susy breaking and can explain the origin of the large hierarchy between the planck ( string , grand unified ( gut ) ) scale and the weak scale . the gmsb itself , however , has nothing to say about the @xmath0-parameter unless one introduces extra fields which couple to the particle content of the mssm . the @xmath0-problem in the gmsb is the primary interest of this paper . many solutions to the @xmath0-problem have been suggested by different authors and all of them require the introduction of new fields and/or interactions . some of these solutions will be reviewed shortly . in the original models @xcite , susy is broken dynamically in a so - called susy - breaking sector and the breaking effects are transmitted to the supersymmetric standard model via a messenger sector . the energy scale of the messenger sector is given by @xmath26@xmath27 gev . there are , however , models which do not have a separate messenger sector so that the sector which breaks susy dynamically is directly coupled to the standard model gauge group @xcite . in this case , the effective messenger scale tends to be much higher . for our purposes it is enough to employ a simple version of the messenger sector , as in the original models , and take the messenger scale @xmath28 as a free parameter . the messenger sector can be described by the superpotential @xmath29 where @xmath30 is a singlet superfield , @xmath31 are charged under a @xmath32 associated with the susy - breaking sector and are singlets under the standard model @xmath33 gauge group . the superfield @xmath34 ( @xmath35 ) transforms as a @xmath36 under the standard model , while @xmath37 ( @xmath38 ) transforms as @xmath39 . we assume that the scalar components of @xmath31 acquire negative susy - breaking masses - squared due to its interaction with the susy - breaking sector ( usually accomplished by the so - called `` messenger @xmath32 '' gauge interaction @xcite ) , and the potential associated with the scalar component of @xmath30 reads @xmath40 neglecting terms containing @xmath37 or @xmath34 . it is easy to see that the scalar and the @xmath41 components of @xmath30 acquire vacuum expectation values ( vevs ) @xmath42 and @xmath43 and therefore @xmath34 and @xmath37 acquire supersymmetric masses proportional to @xmath44 and susy - breaking masses - squared proportional to @xmath43 . , @xmath45 in this potential @xcite . this problem can be avoided by introducing more @xmath30 fields to the messenger sector . such details are , however , irrelevant for the rest of our discussion . ] this effect feeds down to the mssm through loop corrections . gauginos acquire majorana masses at one loop , while sfermions acquire susy - breaking masses - squared at two loops . the calculation of these soft susy - breaking parameters was done long ago ( see @xcite ) and its result is well known . at the messenger scale : @xmath46 @xmath47 here and below , all @xmath48 are in the @xmath49 normalization , @xmath50 in the messenger sector discussed above , and @xmath51 determines the number of messenger sector superfields responsible for mediating susy breaking . in the example we described above , which will be referred to as the model with the minimal gmsb , @xmath52 . @xmath53 is the hypercharge of the particle , @xmath54 for weak @xmath55 doublets ( zero for singlets ) and @xmath56 for color triplets ( zero for color singlets ) . ( [ sfermionmass ] ) guarantees that squarks of different families are degenerate at the messenger scale and therefore fcnc effects are safely suppressed . it is interesting to note that , for small @xmath51 , gaugino masses and sfermion masses are comparable . for very large @xmath51 , on the other hand , sfermion masses can be significantly smaller than gaugino masses ( by a factor @xmath57 ) . in the mechanism described above , trilinear couplings are not generated at the same order ( in loop expansion ) at the messenger scale . this is not the case in general , and some models can generate trilinear couplings with values comparable to the other soft susy - breaking parameters even at the messenger scale @xcite . we will , for most of our discussions , consider @xmath58 unless otherwise noted . the gmsb does not generate a @xmath0-term because of the non - renormalization theorem . therefore @xmath0 is an input of the model , and , because it has dimensions of mass , its only nonzero natural value is @xmath59 ( @xmath60 , @xmath61 ) . this is clearly not allowed phenomenologically . the @xmath0-term must , therefore , be forbidden at the planck scale ( by , say , a @xmath62 symmetry ) and generated dynamically . below , we review various attempts to generate the @xmath0-term in the context of the gmsb . the following list is not meant to be exhaustive and our descriptions of the various attempts are by no means complete . the review below only intends to show that many attempts have been made while none of them appears to be entirely satisfactory . the simplest solution would be to introduce a term in the superpotential @xcite @xmath63 where @xmath30 is the singlet superfield in eq . ( [ messengerw ] ) . in such a scenario @xmath64 and @xmath65 . @xmath66 is the susy - breaking higgs mixing mass - squared in eq . ( [ eq : soft ] ) . phenomenology imposes that both @xmath0 and @xmath67 are of the order of the weak scale , unless one is willing to accept a drastic cancellation among parameters to reproduce the observed @xmath2 . therefore , @xmath68 @xmath69 and @xmath70 this situation is already excluded experimentally . ( [ gauginomass ] ) states that the gluino mass is given by @xmath71 , and if eq . ( [ fs / m=100 ] ) is satisfied one would arrive at @xmath72 gev , which is unacceptable . the same is true for all the other soft susy - breaking masses . this is a general consequence of eq . ( [ mubmu ] ) . it implies that @xmath73 if all experimental bounds on the susy spectrum are to be satisfied , while @xmath74 breaking requires eq . ( [ mu = bmu ] ) . some authors refer to this puzzle as the @xmath0-problem in the gmsb @xcite . another simple solution that does not require the introduction of any extra superfields into the theory couples the higgs superfields to the @xmath34 superfields present in eq . ( [ messengerw ] ) @xcite . in the minimal messenger sector @xcite , one may have , instead of @xmath34 and @xmath37 , a complete @xmath75 multiplet of @xmath49 to preserve the gauge coupling unification . one can also use a @xmath76 for this purpose , and generate gaugino masses and scalar masses - squared with @xmath77 . in this case , one can couple the components @xmath78 in * 10 * that have the same quantum numbers as left - handed quark doublets and the components @xmath79 that have the same quantum numbers as right - handed up quarks ( or their corresponding components in @xmath80 ) to the higgs doublets . explicitly , @xmath81 . this will induce , in the lagrangian , a one - loop term proportional to @xmath82 the @xmath41 vacuum expectation value of @xmath30 will generate @xmath83 and @xmath84 . again one runs into eq . ( [ mubmu ] ) and must hunt for other solutions . all of the models described above couple the mssm higgs superfields to those in the messenger sector . not only did we encounter the problem of eq . ( [ mubmu ] ) , but some of the coupling constants introduced had to be made fairly small because of the magnitude of @xmath44 and @xmath43 . another class of solutions tries to get around this issue by introducing another singlet superfield , whose vacuum expectation value would generate the @xmath0-term . one motivation for such models is to utilize the extra singlet to solve the doublet - triplet higgs splitting in @xmath49 grand unified theories via a sliding singlet mechanism @xcite . this mechanism is known to be unstable against radiative corrections if the soft susy - breaking parameters are generated at a scale higher than the gut scale , but can be stable for the low - energy gmsb @xcite . ciafaloni and pomarol @xcite claim that such a solution would generate a viable @xmath0-term . we believe , however , that the conditions that they impose on the soft susy - breaking parameters can never be satisfied in the context of the gmsb , where all soft susy - breaking masses are tightly related . we will comment on this in section [ sec : mod ] . the simplest model with the addition of an extra singlet one can imagine , referred to as the nmssm @xcite , involves substituting the @xmath0-term in the mssm superpotential by @xmath85 the minimization of the scalar potential for @xmath86 , @xmath87 and @xmath88 at the weak scale should produce vevs @xmath89 and @xmath90 for both higgs bosons , thus breaking @xmath74 , and @xmath91 for the singlet . @xmath0 would be equal to @xmath92 . the @xmath66 term would arise due to renormalization group ( rg ) running of the @xmath93-term @xmath94 from the messenger scale to the weak scale . @xmath66 would be equal to @xmath95 . dine and nelson @xcite claim that this model does not work for the low - energy gmsb . a detailed analysis was not presented in their paper , and we will explain the problem in section [ sec : nmssm ] . they suggest the introduction of an extra light pair of @xmath96 and @xmath97 as a means to produce a viable spectrum . they did not , however , publish a quantitative analysis of the model , and say nothing about its naturalness . agashe and graesser @xcite study this scenario and show that there is indeed a solution , but it is fine - tuned . they present a possibility to ease the fine - tuning by employing many lepton - like messengers while keeping the number of quark - like messengers small . in sections [ sec : nmssm ] and [ sec : mod ] , we analyze in great detail the case for both the high- and low - energy gmsb . there are ways of giving @xmath88 a vev which are not related to electroweak symmetry breaking . in ref . @xcite two mechanisms are introduced , neither of them very appealing , where the @xmath88 vev is generated at the messenger scale . namely , @xmath98 or @xmath99 in addition to the nmssm . @xmath30 , @xmath34 and @xmath37 are the messenger sector superfields present in eq . ( [ messengerw ] ) . in the case of eq . ( [ dns1 ] ) , a potential @xmath100 is generated for @xmath88 in the presence of @xmath42 and @xmath101 vevs . if one assumes @xmath102 to be small , @xmath88 develops a vev @xmath103 , and @xmath104 assuming all other couplings to be of order one . it is easy to see , a posteriori , that @xmath102 must indeed be small if one is to generate a phenomenologically viable @xmath0 . unfortunately this case requires that the soft susy - breaking masses - squared @xmath105 and @xmath106 are accidentally of the same order of magnitude . has to be small , its smallness is natural in the sense of t hooft . it can be interpreted as being generated due to the breaking of some global symmetry , such as @xmath107 and @xmath108 , while @xmath30 is invariant . this type of symmetry would also explain the suppression of a term @xmath109 in the superpotential , which would be of order @xmath110 . ] the superpotential coupling eq . ( [ dns2 ] ) would lead to a potential @xmath111 the linear terms in @xmath88 arise via tadpole one - loop diagrams involving @xmath34 s and @xmath37 s . this would lead to @xmath112 . again @xmath113 and @xmath114 would have to be small . this solution still faces the problem of explaining why a term @xmath115 is not present in the superpotential . note that the presence of such a term would lead to an unacceptably large vev for @xmath88 . one may argue , however , that this is `` technically natural '' because the absence of a term in the superpotential is preserved by radiative corrections . an even more serious problem is the need to suppress the kinetic mixing @xmath116 to ensure @xmath117 ; an unacceptably large @xmath118 would be generated otherwise . an order unity kinetic mixing can be induced via radiative corrections between the ultraviolet cutoff , say the planck scale , and the messenger scale , and the bare parameter has to be chosen very carefully so that the unwanted mixing term can be canceled at the messenger scale . this kinetic mixing can be forbidden if there are two sets of messenger fields and if the field @xmath88 couples off - diagonally , _ e.g. _ , @xmath119 etc @xcite . then the tadpole term mentioned above is also forbidden , but a negative mass squared for the @xmath88 field can be generated instead . this would lead to the nmssm in a successful manner ; again the parameters must be carefully chosen as in the nmssm with extra light quark pairs ( see section 5 ) . another solution with extra singlets , which points an interesting way around eq . ( [ mubmu ] ) , was suggested by dvali , giudice and pomarol @xcite . their idea is to generate the @xmath0-term via the following one - loop effective term in the lagrangian : @xmath120 where @xmath121 is the supersymmetric covariant derivative . this works because @xmath122 cancels @xmath123 in @xmath30 , while leaving @xmath124 in @xmath125 . then the integral over @xmath126 can be done and the @xmath0-term is generated , while @xmath127 is not . the @xmath66 term would arise at higher loops , or via some other mechanism . an explicit realization of this mechanism @xcite is the following . suppose a singlet field @xmath88 acquires a linear term @xmath128 in the superpotential due to its coupling to the messenger sector . then the superpotential @xmath129 leads to a minimum with @xmath130 , @xmath131 . however , by further coupling @xmath88 to the messenger superfields , _ i.e. _ @xmath132 etc , a one - loop diagram of messenger fields generates the operator @xmath133 , which contains @xmath134 . note that this is the same linear potential generated in the case of eq . ( [ dns2 ] ) . this tadpole term induces a vev for @xmath88 of order @xmath135 which is of the order of the weak scale if @xmath136 . the @xmath53 field plays a crucial role : it _ slides _ to cancel the @xmath41-component vev of @xmath88 before the tadpole is added and , after susy is broken , its vev is shifted and leads to @xmath137 , as required by phenomenology . note that the @xmath0-parameter obtained here can be understood as a consequence of the effective lagrangian eq . ( [ dgpddr ] ) , which is generated upon integrating out @xmath88 and @xmath53 before substituting the effect of the vevs of @xmath30 . the necessary linear term ( @xmath138 ) in the superpotential for @xmath88 can be easily generated by the kinetic mixing between @xmath88 and @xmath30 or also by other mechanisms , as pointed out in reference @xcite . one apparent drawback of this realization is that one needs a set of new fields whose interactions are arranged in a rather special way . furthermore one would expect the presence of a term proportional to @xmath139 in the superpotential . this happens because both @xmath30 and @xmath88 couple to the messengers , that is , @xmath140 , and have , therefore , the same quantum numbers . we have already argued that a coupling @xmath139 has to vanish ( see eq . ( [ simple ] ) ) . finally we point out that this model also suffers from the cancellation problem present in the mssm ( see section 3 ) . dine , nelson , nir and shirman @xcite suggest yet another way of generating a @xmath0-term with the introduction of an extra singlet . it was inspired by flavor symmetry models in @xcite , and resembles a modified version of the nmssm + eq . ( [ dns1 ] ) : @xmath141 where @xmath142 is the planck mass . when @xmath143 , @xmath52 and @xmath144 , it is easy to check that @xmath145 . we assume the other couplings to be of order 1 . it is also easy to see that one would require a very small , carefully chosen coupling @xmath146 in order to guarantee @xmath147 gev . it is worth noting that this mechanism does not generate an @xmath66 term . at last we would like to mention another interesting possibility , pointed out by yanagida @xcite and nilles and polonsky @xcite . their models utilize the accidental equality @xmath148 , where @xmath149 gev is the scale of dynamical susy breaking ( dsb ) in models with the low - energy gmsb and @xmath150 the reduced planck mass . by introducing a new susy - preserving sector with strong gauge dynamics , yanagida s model generates a vev for the superpotential which cancels the cosmological constant from the dsb sector . the constant superpotential in turn generates a @xmath0-term of order @xmath151 tev . the phenomenology of this model is the same as the previous one ( see eq . ( [ flavorinsp ] ) ) . the model by nilles and polonsky makes use of the planck - scale suppressed khler potential , @xmath152 , where @xmath153 is a chiral superfield in the dsb sector with an @xmath41-component vev . this operator may be present at the tree - level , but may also be generated by gravitational effects . it generates a tadpole for the singlet @xmath88 : @xmath154 . together with the @xmath155 superpotential of the nmssm , it generates a vev for @xmath88 of order @xmath156 . even though these models generate the correct @xmath0-term of order the weak scale in the models with the low - energy gmsb , this would not work for the high - energy gmsb . we consider that none of the mechanisms outlined above are entirely satisfactory . most of them require a very specific choice of parameters and the introduction of extra matter at or slightly above the weak scale . furthermore , most of them have not been studied quantitatively ( see , however , ref . @xcite ) , and there is no guarantee that they indeed generate the correct electroweak symmetry breaking pattern and an experimentally viable spectrum . and last , but not least , there is no study of how natural such a solution is , given that a viable pattern of electroweak symmetry breaking can be generated . it is , therefore , part of our goal to study the simplest of the models mentioned above in detail . before that , we would like to review the status of electroweak symmetry breaking in the mssm , where the @xmath0-term is introduced `` by hand . '' we will point out that , in the case of the gmsb , the current lower bounds on superparticle masses already require an order @xmath1% cancellation between the @xmath0-parameter and the soft susy - breaking parameters . we reviewed various proposals to generate the @xmath0-parameter in models with the gmsb . in this section , we review how electroweak symmetry breaking occurs in the mssm , assuming that the @xmath0-parameter and @xmath66 are somehow generated . in particular , we point out a need for an order @xmath1% cancellation between @xmath0-parameter and soft susy - breaking parameters in models with the gmsb given the current experimental lower bounds on superparticle masses . note that the case of the nmssm is different because the @xmath0-parameter is generated together with electroweak symmetry breaking and hence the two problems can not be clearly separated . this will be discussed in the next two sections . the tree - level higgs potential in the mssm is given by @xmath157 where the mass parameters involve both the supersymmetric @xmath0-term and the soft susy - breaking terms , @xmath158 in the mssm , one can show that the vacuum can always be gauge rotated to the following configuration @xmath159 the two expectation values need to satisfy @xmath160 in order to reproduce the observed @xmath5 , and it is conventional to parametrize them by @xmath161 , @xmath162 . the minimization condition of the potential can be rewritten in the following form : @xmath163 it has been claimed that electroweak symmetry breaking is natural in the mssm because @xmath164 is easily driven negative due to the presence the top yukawa coupling in its rg evolution . in models with the minimal gmsb such as the original ones in @xcite , the boundary condition for the supersymmetry breaking parameters are given by eqs . ( [ gauginomass ] , [ sfermionmass ] , [ aterms ] ) . a simple one - loop approximation is valid in the case of the low - energy gmsb because of the small logarithm between the messenger scale @xmath28 and the electroweak scale , and one finds @xmath165 which is always negative . the need for a cancellation between the @xmath0-parameter and soft susy - breaking masses can be seen as follows . experimental constraints bound the superparticle masses from below , which hence set a lower limit for the ratio @xmath166 . therefore one finds that @xmath167 is bounded from below . on the other hand , in order for the observed @xmath168 to be reproduced , the @xmath0-parameter is constrained by eq . ( [ eq : mueq ] ) . for a moderately large @xmath169 , @xmath170 can be completely neglected and one finds @xmath171 this equation requires a cancellation between @xmath172 and ( negative ) @xmath164 to reproduce @xmath173 correctly . the degree of cancellation is given by @xmath174 . ) is free to vary before a given output parameter ( in this case @xmath168 ) changes significantly . explicitly , the degree of cancellation is @xmath175 . this definition corresponds to the inverse of the barbieri giudice function @xcite . ] to determine the lower limit on @xmath167 , we consider a number of experimental constraints @xcite . one is that the gluino must be heavier than 190 gev , which becomes stronger if the squarks have comparable masses . the second is that the right - handed selectron must be heavier than 80 gev . turns out to be large , it is _ a posteriori _ justified to assume that the lightest neutralino is almost pure bino . then the gmsb predicts the relation between selectron and the bino masses , and hence we have a fairly reliable lower bound . ] for large @xmath176 , the right - handed stau may become rather light ; we then require @xmath177 gev if it decays into tau and a neutralino or gravitino , and @xmath178 gev if it does not decay inside the detector . we also considered the lightest chargino to be heavier than 63 gev . the most recent lower bound on the chargino mass @xcite is @xmath179 gev , which leaves our analysis virtually unchanged . let us first discuss the case of the minimal low - energy gmsb with small @xmath176 to make the argument clear . here we consider @xmath180 gev . the gluino mass constraint requires @xmath181 this bound itself is independent from the messenger scale . however , the gluino mass bound depends on the mass of the squarks , and it strengthens if the squark masses are comparable to the gluino mass . for the minimal low - energy gmsb , squarks are significantly heavier than the gluino and we can use the bound above . a more stringent constraint is derived from the requirement that the right - handed sleptons are heavier than 80 gev . including the one - loop renormalization group evolution and the @xmath182-term , we find @xmath183 therefore we find @xmath184 for the most conservative case @xmath185 . with this lower bound we find @xmath186 using the one - loop running of @xmath164 , we obtain @xmath187 and as a result of the minimization condition , @xmath188 this requires a cancellation of 7% in order to obtain the correct @xmath168 . even though this level of cancellation is not of immediate concern , this analysis shows the need for a certain amount of cancellation which will become worse as experimental lower bounds on superparticle masses improve . = 0.6 as it is clear from the argument above , the actual lower bound on @xmath0 depends on the messenger scale and @xmath176 . we have studied this issue numerically using the experimental bounds quoted above and found the lowest possible value of @xmath0 as a function of the messenger scale . in fig . [ fig : mumin](a ) we present bounds for three values of @xmath176 . the lower bound on @xmath0 comes from one of the various experimental constraints . for instance , the @xmath189 case is dominated by the lower bound on @xmath190 up to a messenger scale of @xmath191 gev , after which the gluino mass bound is more important.=1 ) . for larger @xmath51 the gluino bound becomes less important and the slepton bounds dominate up to the gut scale . ] the case of @xmath192 has a similar behavior . the situation is more complex and interesting for @xmath193 . for a messenger scale of up to @xmath194 gev , the stau is the lightest supersymmetric particle ( except for the gravitino ) . it decays inside the detector to tau and gravitino for the lowest messenger scale , but leaves the detector without decaying for higher messenger scales . this stable stau provides the strongest constraint . from messenger scales above @xmath195 gev the stau decays inside the detector to tau and neutralino . this bound dominates up to @xmath196 gev , when the gluino bound dominates . the chargino bound is comparable to that of the gluino for the gut scale ( @xmath197 gev ) . in fig . [ fig : mumin](b ) we show the minimum value of @xmath0 as a function of @xmath176 for a fixed messenger scale ( @xmath198 gev ) . the @xmath176 dependence can be easily understood as follows . starting from low @xmath176 , increasing @xmath176 decreases the top yukawa coupling , and hence @xmath199 receives a less negative contribution from the top - stop loop . therefore a lower value of @xmath0 is allowed . this part is dominated by the @xmath200 bound . however beyond @xmath201 , the bottom and tau yukawa coupling become important . in fact , the scalar tau mass is pushed down both because of the loop effect and left - right mixing , and the experimental lower bound on @xmath202 becomes stronger . beyond @xmath203 , the stau does not decay inside the detector for this choice of the messenger scale and the constraint is even more stringent . this in turn leads to a more negative @xmath204 and hence a larger @xmath0 . combining both the messenger scale dependence and @xmath176 dependence , we conclude that the most conservative current limit is @xmath205 the required cancellation between @xmath172 and soft susy - breaking parameters in order to reproduce the observed @xmath5 is @xmath206% . note that this level of cancellation is the absolute minimum , and a more accurate cancellation is required for most of the parameter space . in the case of minimal supergravity models , where all scalars have the universal susy - breaking mass - squared @xmath207 , all gauginos have mass @xmath208 and all @xmath93-terms are given by @xmath209 for @xmath210 , at the gut scale , the situation appears to be much better . the renormalization group equations can be solved numerically for each choice of @xmath176 . as an example we take @xmath189 and find @xmath211 by requiring @xmath212 gev and @xmath213 gev ( this is a rough bound inferred from the gluino bound @xmath214 gev ) , we find @xmath215 gev . this basically does not require any cancellation , since @xmath216 . the situation can be somewhat ameliorated in the mssm if there is a fayet illioupoulos @xmath182-term for the @xmath217 gauge group . such a @xmath182-term is known to arise in many ways , such as kinetic mixing of the @xmath217 and @xmath218 gauge fields @xcite . the running of all the parameters remains the same except that one adds another contribution from the @xmath219 at the weak scale . if the sign is appropriate , it increases @xmath220 and @xmath221 ( less negative ) while decreasing @xmath222 . all of these help push the parameters relevant for electroweak symmetry breaking in the right direction . larger @xmath223 reduces the lower bound on @xmath224 , and a less negative @xmath164 is also welcome . therefore the sensitivity to @xmath0 ( required cancellation between @xmath0 and soft susy - breaking parameters ) in the mssm can be improved in the presence of a @xmath219 with the appropriate sign . we will see in the next two sections that the situation in the nmssm is much worse . there is no phenomenologically viable solution to electroweak symmetry breaking . one can modify the model to generate a large negative mass - squared for the singlet field and then find a viable solution . this solution also requires a cancellation among parameters which has the same origin as the cancellation present in the mssm . we will also see that the addition of the fayet does not improve the situation within the nmssm . in this section we study the feasibility of implementing the gmsb in the framework of the next - to - minimal supersymmetric standard model ( nmssm ) . we begin our presentation by introducing the nmssm : its particle content , superpotential , and soft susy - breaking terms . we briefly review the major steps in our analysis : the boundary conditions for the breaking terms , the rg evolution , and the minimization of the weak - scale one - loop effective potential . we then describe the results of a numerical scan of a large portion of the model s parameter space . we find that it is impossible to evade the present - day experimental constraints . we further strengthen this argument by providing a semi - analytical explanation for the inevitability of this conclusion . the nmssm represents an attempt to solve the @xmath0-problem of the mssm in the simplest and most direct way : the spectrum of the mssm is augmented by a gauge singlet superfield _ n _ , which couples to @xmath225 and plays the role of the @xmath0-term once it develops a nonzero vacuum expectation value @xcite . the original @xmath0-term is banned from the theory so that there are no dimensionful parameters left in the superpotential . the vev of the scalar component of _ n _ is determined by minimizing the scalar potential with respect to @xmath16 , @xmath15 , and @xmath88 simultaneously . it is natural to expect the vevs to be of the same order of magnitude for all three fields , thus generating an effective @xmath0-parameter of order the weak scale , as required by phenomenology . the complex scalar _ n _ introduces two additional degrees of freedom to the higgs sector . therefore , the particle spectrum of the nmssm contains three cp even higgs scalars , two cp odd higgs scalars , and one charged higgs scalar . immediately , there is a problem : one of the pseudoscalar higgs bosons is massless . this happens because the superpotential @xmath226 has a peccei - quinn symmetry @xmath227 . this symmetry is spontaneously broken by the vevs of the fields , making one of the pseudoscalars massless . the standard solution to this problem is to introduce a term cubic in _ n _ , which explicitly breaks the symmetry mentioned above . this term is allowed by the gauge symmetries of the model and does not contain a dimensionful coupling constant , so it is generically expected to be present in the superpotential . one , however , still has to worry about a light pseudoscalar higgs boson . as we will show shortly , its mass can also be small because of the presence of a different ( approximate ) @xmath32 symmetry . overall , the only change made to the mssm superpotential is the following : @xmath228 while the corresponding change to the soft susy - breaking part of the potential is : @xmath229 one can determine the vevs of the higgs fields @xmath16 , @xmath15 , and @xmath88 by minimizing the scalar potential , which at the tree - level consists of the @xmath41-terms , @xmath182-terms , and soft susy - breaking terms : @xmath230 an important fact to notice is that both @xmath231 and @xmath232 remain unchanged when @xmath233 are all rotated by the same phase . in fact , only the soft susy - breaking @xmath93-terms are not invariant under this transformation . this can be potentially dangerous , because we , in general , consider the @xmath93-terms to be zero at the messenger scale , and their sizes at the weak scale are determined by the rg evolution . if the generated values of @xmath234 and @xmath235 are not large enough , our scalar potential has an approximate @xmath32 symmetry . this symmetry is spontaneously broken by the vacuum expectation values of the higgs fields , and , as before , we have to worry about a light pseudoscalar higgs boson . we denote the vevs of the neutral components of the higgs fields by @xmath236 and @xmath237 , as in section [ mssmsection ] , and the vev of the singlet field by _ x _ : @xmath238 as a function of these vevs , the potential has the form @xmath239 it is well known that some of the higgs boson masses receive significant contributions from radiative corrections . in our numerical analysis we account for that by employing the one - loop effective potential @xmath240 in this expression @xmath241 is a field - dependent scalar mass - squared matrix , and @xmath0 is the @xmath242 renormalization scale . as we have indicated explicitly , the values of the various parameters entering @xmath243 depend on the choice of this scale . to the leading order this dependence is canceled when the second term on the right hand side of eq . ( [ oneloopeffective ] ) is included , and the result of minimizing @xmath244 is less sensitive to the choice of the scale where one stops running the rg equations . ( canceling out this dependence completely would require calculating radiative corrections to all orders . ) the matrix @xmath245 depends on the field vevs @xmath246 through the yukawa couplings of the higgs fields to various other particles . what plays a crucial role here is not the absolute values of the masses , but rather the rate of their change as one changes @xmath246 . therefore , the most important contribution comes from the field - dependent masses of the top quark and squarks , which have the largest yukawa coupling . denoting their mass eigenvalues by @xmath247 , @xmath248 and @xmath249 respectively , the contribution to @xmath244 from radiative corrections due to these states is @xmath250.\end{aligned}\ ] ] in models with the gmsb the values of the soft susy - breaking terms are specified at the messenger scale by eqs . ( [ gauginomass ] ) , ( [ sfermionmass ] ) and ( [ aterms ] ) . their values at the weak scale can be determined by solving the rg equations given in [ appendixa ] . the model has five input parameters : @xmath251 , @xmath252 , @xmath253 , @xmath202 , and @xmath51 . ( note that the only dimensionful input parameter is @xmath202 , and its magnitude will determine the overall scale of the vevs and the soft susy - breaking masses . ) there are , however , two constrains which must be satisfied at the weak scale : @xmath254 gev and @xmath255 gev .. ] a common approach is to use the minimization conditions and rg equations to solve for the inputs , given a phenomenologically allowed set of weak scale outputs . in the case of a high messenger scale , however , no easily invertible solution for the rg equations is available . instead , we simply choose to tackle the problem numerically . after running down the rg equations and minimizing the higgs potential once , we iterate this procedure , each time adjusting the value of the parameter _ b _ to fix the overall scale of the vevs and masses , while simultaneously changing the dimensionless couplings to correctly reproduce the top quark mass . this iteration process , in fact , converges fairly quickly . using the procedure above , we perform a numerical scan of a large portion of the parameter space . we study the low - energy particle spectrum for various messenger scales @xmath28 , numbers of messengers _ n _ , and values of the couplings @xmath256 . it is interesting to note that it is very easy to generate non - zero vevs for @xmath257 and @xmath88 , even when @xmath258 is a small _ positive _ number . this is because the terms @xmath259 , when @xmath260 and @xmath261 are of the same sign , both `` push '' the vev of the real component of the singlet away from the origin . unfortunately , we find that , for any choice of values of the input parameters , there are always particles with unacceptably small masses . to illustrate the situation , we present in table [ table1 ] our numerical results for several representative points in different `` corners '' of the parameter space . the first two points represent the typical situation for the case of the low - energy gmsb , the next two are representative of the case of the high - energy gmsb , and the last one explores the extreme case of @xmath262 gev . points 1 and 3 have relatively large values of @xmath253 , while points 2 and 4 have @xmath263 . notice that in the table we did not consider a similar limiting case for @xmath252 . this is not a coincidence . it turns out that , for @xmath264 , the dominant term in the potential is @xmath232 , and @xmath176 is forced to values very close to one . in this case , in order to correctly reproduce the top quark mass , one is forced to choose @xmath265 at the weak scale such that @xmath265 hits the landau pole below the gut scale . we have chosen to list only the cases where the couplings in the superpotential remain perturbative up to the gut scale . we make , however , no such assumption in our analysis in the next subsection . .the numerically determined nmssm parameters for five sample points in the parameter space . here @xmath266 and @xmath267 refer to the eigenvalues of the scalar and pseudoscalar higgs mass matrices respectively , and @xmath268 denotes the mass of the right - handed selectron . the values of @xmath252 , @xmath253 , and @xmath265 are given at the weak scale . all the other quantities have been defined earlier in the text . [ cols="^,^,^,^,^,^,^ " , ] it can be seen that , in all the cases presented in table [ table1 ] , there are particles with unacceptably small masses . the result for the low - energy gmsb is not new and has been known for several years @xcite . on the other hand , the situation with a high messenger scale had not been quantitatively studied in the literature to date . one expected feature that we indeed see in points 3 and 5 is the increase of the pseudoscalar higgs boson mass with @xmath28 . this happens because the magnitude of @xmath234 , generated by running the rg equations , increases with the messenger scale , and it is @xmath234 that breaks the @xmath32 symmetry of the potential , as discussed before . another result that could have been anticipated is the smallness of the mass of the light pseudoscalar higgs when @xmath269 ( points 2 and 4 ) . this is due to the peccei - quinn symmetry , which is restored in this limit . what is surprising is that raising the messenger scale by 10 orders of magnitude does not bring any other significant changes to the particle spectrum . the masses of the gluino , right - handed selectron , and scalar higgs boson still remain small . in this subsection we present a rather simple semi - analytical argument which explains why there can be no phenomenologically acceptable solution to the nmssm with the gmsb . we show that if one assumes that such a solution exists , one arrives at a contradiction . we also explain some of the features of the numerical solutions presented in the previous subsection . suppose that for some point in the parameter space an acceptable solution exists . the problem to be addressed is the smallness of the selectron , gluino , and higgs masses . we choose to base our analysis on the right - handed selectron mass constraint . the magnitude of @xmath268 is directly proportional to the size of the @xmath202-parameter . in our numerical procedure the value of @xmath202 is chosen in such a way that @xmath270 gev . a typical value of @xmath202 obtained in this way yields a very small selectron mass ( @xmath271 gev ) , gluino mass ( @xmath272 gev ) , and soft susy - breaking masses for the higgs bosons ( @xmath273 gev@xmath274 , @xmath275 gev@xmath274 ) . it is , therefore , obvious that the only chance of obtaining an acceptable value for the selectron mass is to raise _ b _ , by a factor of three or more , and try to arrange the other parameters in such a way that @xmath276 remains 174 gev . since _ b _ feeds into all soft susy - breaking masses , their absolute values will also increase . for example , imposing @xmath277 gev forces @xmath278 for a messenger scale of @xmath7 gev . for different messenger scales the bound becomes even more stringent , as shown in fig [ fig : m2bounds](a ) . = 0.8 to determine the consequences of raising the soft susy - breaking masses , we analyze the higgs potential ( eq . ( [ treepotential ] ) ) . the extremization conditions at tree level are @xmath279 the first two equations ( eqs . ( [ minim_conditions_v1 ] ) and ( [ minim_conditions_v2 ] ) ) closely resemble the corresponding ones in the mssm case . in fact , the only difference in the nmssm is the presence of the first term on the right - hand side of eq . ( [ minim_conditions_v1 ] ) and eq . ( [ minim_conditions_v2 ] ) . this term originates from @xmath280 and is , therefore , absent in the mssm . dividing eq . ( [ minim_conditions_v1 ] ) by @xmath237 , eq . ( [ minim_conditions_v2 ] ) by @xmath236 , and subtracting the two expressions , we can cancel out this term . as a result , we obtain @xmath281 note that this equation is identical to eq . ( [ eq : mueq ] ) with @xmath282 . to obtain the nmssm analog of eq . ( [ eq : m3eq ] ) we divide eq . ( [ minim_conditions_v1 ] ) by @xmath236 , eq . ( [ minim_conditions_v2 ] ) by @xmath237 and _ add _ the two . solving for @xmath283 ( @xmath284 ) we find : @xmath285 eqs . ( [ eq : m3eq ] ) and ( [ bmu ] ) differ only by the contribution from @xmath286 . ( [ mu^2 ] ) states that the value of the effective @xmath0-parameter generated in this model is subject to a rather stringent bound : @xmath287 , which , if one imposes @xmath288 , translates into @xmath289 . notice that the origin of this bound is the same as of the bound on the size of the @xmath0-parameter derived in section [ mssmsection ] , since the condition given by eq . ( [ mu^2 ] ) is the same in both cases . in the present case , however , the bound is stronger because @xmath176 is no longer a free parameter but is determined by minimizing the higgs potential . so far we have only looked at the first two extremization conditions . we now turn our attention to eq . ( [ minim_conditions_x ] ) . solving for @xmath290 in eq . ( [ mu^2 ] ) , one can rewrite eq . ( [ minim_conditions_x ] ) as @xmath291 while we have shown that phenomenology requires the expression in parenthesis on the left - hand side to be larger than @xmath292 , the terms on the right - hand side are all much smaller , because @xmath293 are zero at the messenger scale and the effects of the rg running are relatively small ( see fig . [ fig : m2bounds ] ) . this means that the above equation can never be satisfied unless @xmath294 . an immediate consequence of the @xmath295 limit is that the mass of the lightest pseudoscalar higgs goes to zero , as it becomes a nambu - goldstone boson . ( it is for this reason that _ k _ was introduced in the first place . ) furthermore in the limit of large @xmath0 and small @xmath253 the determinant of the scalar higgs mass - squared matrix becomes negative , which means that the extremum point given by eqs . ( [ minim_conditions_v1]-[minim_conditions_x ] ) ceases to be a minimum . to show this we first derive a relationship between _ k _ and @xmath296 . that relationship can be derived from eq . ( [ mu^2 ] ) and eq . ( [ bmu ] ) . neglecting @xmath297 , @xmath283 , and @xmath298 in comparison to @xmath299 and @xmath199 , we find that @xmath300 here @xmath301 . ) it was necessary to assume that @xmath302 . this translates into two requirements : @xmath303 . we conclude that for large soft susy - breaking higgs masses - squared it is necessary to have @xmath304 . ] for @xmath305 , eq . ( [ sin2beta ] ) reduces to @xmath306\ , .\end{aligned}\ ] ] equipped with the last result , we consider the determinant of the scalar higgs mass - squared matrix . the full expression for it is given in [ 3x3matrix ] ; here we only need to identify the leading terms . we are interested in the case @xmath307 , and , as we have argued before , the soft trilinear couplings @xmath234 and @xmath235 can be neglected . for this reason , the dominant terms will be the ones containing the highest power of @xmath0 : @xmath308 that these are indeed the largest terms was checked numerically . taking into account the fact that @xmath261 and @xmath309 are proportional to each other for small @xmath253 , one can easily see that , in the limit @xmath310 , the first term dominates and the determinant is negative . there is no ambiguity with sign redefinitions of @xmath252 or @xmath253 in eq . ( [ determ2 ] ) . ] this completes our argument , and we are now able to state that there can be no phenomenologically viable solution in the context of the nmssm . we could have also based our argument on the gluino mass bound . the experimental constraint @xmath311 gev translates into the requirement @xmath312 ( assuming @xmath52 ) , and the rest of the argument follows unchanged . notice , however , that the bound on @xmath313 weakens if the number of messenger fields is taken to be very large . we now turn to the issue of interpreting the numerical results of the previous subsection . we would like to understand , for instance , why the values of the singlet vev @xmath91 in table [ table1 ] are always smaller than the vevs of the higgs doublets and , furthermore , why @xmath91 is only several gev for a low messenger scale . the answer comes from considering the extremization condition for @xmath91 : @xmath314 where we omitted the terms @xmath258 and @xmath315 ( @xmath316 for all the points in the table ) . for most of the parameter space the cubic term in @xmath91 can also be neglected , giving @xmath317 thus the smallness of @xmath91 is related to the fact that @xmath234 is small . the above approximation holds as long as @xmath318 which is not satisfied only for point 5 in table [ table1 ] . for point 5 the value of @xmath91 can be approximated by @xmath319 again @xmath320 and therefore @xmath321 gev . knowing that @xmath91 is small in this model we can derive another interesting relation . neglecting all the terms containing @xmath91 in eq . ( [ bmu ] ) , we obtain : @xmath322 this explains why the values of the soft susy - breaking masses for the higgs bosons are so similar for very different values of the messenger scale . finally , we can say a few words about the scalar higgs boson masses . in the limit of small @xmath91 ( and hence small @xmath0 ) , the dominant term in the determinant of the scalar higgs mass - squared matrix ( see [ appendixb ] ) is @xmath323 taking into account the fact that , for small @xmath91 , @xmath324 ( see eq . ( [ xa / l ] ) ) , the equation above gives : @xmath325 this explains why changes in the messenger scale have almost no effect on the product of the scalar higgs boson masses ( see table [ table1 ] ) , as long as @xmath252 is unchanged . for @xmath326 , which is what we typically find in this case , eq . ( [ higgsestimate ] ) gives a geometrical average " value of the scalar higgs boson mass of only about 50 gev . this means that , as long as @xmath91 is small , the model necessarily yields phenomenologically unacceptable higgs boson masses . in this section we reexamine the expressions derived in section [ sec : nmssm ] and attempt to modify the nmssm to make it phenomenologically viable . we study several possibilities and comment on the problems that arise . overall , we find none of these possibilities entirely satisfactory . we want to modify the nmssm in a way that allows one to avoid the conclusions of section [ sec : nmssm ] . recall that the crucial step in our analysis there was the observation that eq . ( [ modified_x ] ) could not be satisfied : the left - hand side was always greater than the right - hand side . to obtain a consistent solution one has to somehow make both sides equal . one possibility is to make @xmath258 of the same order of magnitude ( and sign ) as @xmath199 . that could be accomplished by coupling the singlet to some new fields and arranging the parameters in such a way that the susy - breaking mass - squared of the singlet is driven sufficiently negative . this idea was first proposed by dine and nelson in ref . @xcite , who introduced new color - triplet fields @xmath327 and @xmath328 and coupled them to _ n_. the corresponding superpotential is @xmath329 according to eq . ( [ sfermionmass ] ) , the scalar components of @xmath327 and @xmath328 acquire large susy - breaking masses , which can drive @xmath258 sufficiently negative . agashe and graesser in ref . @xcite did a quantitative study of this scenario for the case of the low - energy gmsb . they showed that it is indeed possible to generate a large negative @xmath258 , in the range @xmath330 to @xmath331 , and further demonstrated that , with @xmath258 of this magnitude , one can choose the input parameters in such a way that @xmath270 gev and all experimental constraints are evaded . they also pointed out that in this scenario the input parameters need to be fine - tuned in order to reproduce the above value of @xmath332 . in what follows we give a set of input parameters that yields an acceptable particle spectrum , and then proceed to analyze the sensitivity of the higgs boson vevs to the nmssm coupling constants . we clarify the origin of this sensitivity and also extend the analysis to the case of the high - energy gmsb . as an example of an allowed solution , we consider the case of the low - energy gmsb with @xmath333 tev , @xmath52 , and @xmath334 tev . for @xmath335 , to correctly reproduce @xmath336 we take @xmath251=0.99 , @xmath337 and @xmath252=0.11 at the weak scale . we find that @xmath176 equals @xmath338 for this point . because the magnitude of the product @xmath339 is now quite large , the masses of the gluino and right handed selectron are safe : @xmath340 477 gev , @xmath34193 gev . the vacuum expectation value of the singlet is also large , @xmath91=2.97 tev , which , as was argued earlier , is required by eq . ( [ mu^2 ] ) . the eigenvalues of the scalar higgs mass matrix are 404 , 270 , and 90 gev , and those of the pseudoscalar higgs mass matrix are 400 and 6.7 gev . the last number appears alarmingly small at first sight but , as shown in ref . @xcite , has not been excluded . the reason is that the corresponding eigenstate @xmath342 is almost a pure singlet : @xmath343 the quantitative criterion given in ref . @xcite , based on the constraint from the @xmath344 decay , is @xmath345 and for the parameter set above the left - hand side equals 0.15 . = 0.6 = 0.6 in this scheme it is , therefore , possible to find a point in the parameter space which leads to a phenomenologically viable solution . unfortunately , as we already mentioned , this solution is very sensitive to the choice of the superpotential coupling constants @xmath252 and @xmath253 . in the remainder of this subsection we discuss this issue in detail . the values of the parameters for the set that we have just described had to be chosen in such a way that the top quark and @xmath346boson masses were fixed at their known experimental values . it is interesting to investigate what values of @xmath2 would be predicted for a generic choice of the parameters . in fig . [ fig:3dplot ] we plot the magnitude of the quantity @xmath347 as a function of @xmath252 and @xmath253 . the figure shows that small changes in both @xmath252 and @xmath253 lead to large changes in @xmath332 . this is very similar to the situation in the mssm which was considered in section [ mssmsection ] . there we showed that the value of the @xmath0-parameter had to be chosen very carefully in order to yield the correct value of @xmath332 . in the present case , the points in the parameter space that correspond to values of @xmath332 around 174 gev lie in a very thin band on the @xmath348 plane . also notice that , for this range of @xmath252 and @xmath253 , the slope is the steepest . ( see [ commentonprob ] for comments on this point . ) it is possible to perform the same type of analysis for a higher messenger scale . the same problem is found in that case as well . in fig . [ fig:2plots ] we plot the dependence of @xmath332 on @xmath349 for fixed values of @xmath253 . for comparison , the curve for @xmath350 gev is plotted next to the curve for @xmath351 tev . from the slopes of these curves one can determine the degree of sensitivity with respect to @xmath252 , using the definition in section [ mssmsection ] . the degree of sensitivity , given by @xmath352 , is @xmath353 for the low - energy curve and @xmath8 for the high - energy curve . our numerical results agree with those in ref . @xcite for the low - energy gmsb if the same inputs parameters are used . in order to understand this behavior , we once again turn to the extremization conditions eqs . ( [ minim_conditions_v1]-[minim_conditions_x ] ) . first , we present some qualitative observations . recall that phenomenology requires @xmath354 to be rather large ( of the order @xmath355 tev ) , while @xmath332 has to remain `` small '' ( @xmath270 gev ) to correctly reproduce @xmath5 . as a result , the terms containing high powers of @xmath91 and the terms containing @xmath356 ( @xmath357 ) dominate , while the terms with @xmath237 and @xmath236 are not fixed , and have to absorb the residual difference between the dominant terms . therefore , small percentile changes in the dominant terms can result in large percentile changes in the higgs boson vevs . this is to be contrasted with the situation in the previous section , where @xmath358 was tied to the value of the sum @xmath359 ( see eq . ( [ lambdax ] ) ) . next , we try to identify the main source of this sensitivity . we first consider the dependence of @xmath332 on @xmath252 for fixed @xmath202 , @xmath253 , and @xmath251 . one can use eq . ( [ mu^2 ] ) to solve for @xmath360 and then isolate the largest contribution to @xmath361 . @xmath362 \ , .\end{aligned}\ ] ] using the data that led to fig . [ fig:3dplot ] , we numerically evaluate the derivative around the point @xmath363 . the following are the results of evaluating each of the terms on the right - hand side , respectively : @xmath364 ( gev@xmath274 ) . the largest term is the first one , the next two terms combined provide a 45% correction , and the derivatives of the soft susy - breaking masses can be completely neglected . in [ estimatingsensitivity ] we show how these numbers can be understood by studying the minimization conditions . the fact that the dominant contribution to @xmath365 comes from the first term in eq . ( [ variation ] ) has a very important implication . it means that the problems of cancellation in the nmssm and the mssm are not merely similar , but have _ exactly _ the same origin . indeed , eq . ( [ mu^2 ] ) is the same as eq . ( [ eq : mueq ] ) , and , because in the nmssm @xmath366 depends on @xmath252 mainly through the combination @xmath92 , which plays the role of the @xmath0term , the two models require roughly the same degree of cancellation . the degree of cancellation quoted in section [ mssmsection ] for the mssm is most conservatively 16% , but this is so because one can choose @xmath176 freely in the mssm . on the other hand , @xmath176 is determined by minimizing the potential for the nmssm and can not be chosen arbitrarily to ease the cancellation . for the value of @xmath176 which we obtained in the nmssm , the degree of cancellation is actually comparable ( order a few percent ) in the mssm . the small difference between the two models is due to the dependence of @xmath91 and @xmath176 on @xmath252 . we have discussed the @xmath252 dependence of the higgs boson vev , and now turn to the @xmath253 dependence . [ fig:3dplot ] shows that the points that yield @xmath270 gev form an almost straight line on the @xmath348 plane . it can be shown ( see [ estimatingsensitivity ] ) that in order to keep @xmath332 constant one has to change @xmath253 and @xmath252 according to @xmath367 the sensitivity of @xmath332 to @xmath253 is , thus , related to the sensitivity of @xmath332 to @xmath252 , which , in turn , originates from the need to carefully choose the @xmath0-parameter in the mssm as discussed in section [ mssmsection ] . to summarize , we have shown that this model requires a very particular choice of parameters to yield the correct @xmath346-boson mass . furthermore , we explained that the sensitivity of the @xmath346-boson mass to the nmssm couplings has the same origin as the sensitivity of the @xmath346-boson mass to the value of the @xmath0-parameter in the mssm . we emphasize that the problem is present for both high and low messenger scales , simply because the bound on the @xmath0-parameter does not weaken as one raises the messenger scale . next , we investigate what happens if the @xmath182-term contributions described at the end of section [ mssmsection ] are included . first , we consider the case of the nmssm with no extra particles added . we try to determine if , by introducing the @xmath182-terms , it is possible to make @xmath332 smaller . if that happened , _ could be rescaled back by increasing @xmath202 , and that would raise all masses in the model , as desired . we find that this is not the case . upon adding the @xmath182terms both @xmath368 and @xmath369 change , but @xmath370 , curiously enough , remains virtually constant . this happens because , in the limit @xmath371 , @xmath360 is constrained by eq . ( [ lambdax ] ) , and the change @xmath372 , @xmath373 preserves the quantity @xmath374 . the next question to ask is whether the @xmath182-terms can decrease the degree of cancellation for the case with @xmath327 and @xmath328 added . the answer is again negative and the reason can be seen from eq . ( [ bmu ] ) . recall that the degree of cancellation is controlled by the magnitude of @xmath290 . as long as @xmath375 and @xmath376 can be neglected compared to @xmath359 , eq . ( [ bmu ] ) yields @xmath377 and the relevant quantity is again @xmath374 . at last , we consider the scenario proposed by ciafaloni and pomarol @xcite . they consider a modified version of the nmssm , where @xmath378 , @xmath379 and the value of @xmath234 is large at the messenger scale . their model also contains , in the potential at the weak - scale , a linear term in @xmath88 which is generated by tadpole diagrams and solves the problem of a light pseudoscalar . they find that the requirement of the positivity of the determinant of the scalar higgs boson mass - squared matrix is very restrictive . we repeat part of their analysis to determine if their choice of parameters could indeed lead to a phenomenologically viable electroweak symmetry breaking spectrum . note that , as far as the following is concerned , their model is identical to the nmssm . the full expression for the determinant can be found in [ 3x3matrix ] . in the limit of @xmath310 and @xmath380 @xmath381\ , , \ ] ] where we introduced a variable @xmath382 to conform to the notation used in ref . @xcite . from the extremization conditions for the potential , eqs . ( [ minim_conditions_v1]-[minim_conditions_x ] ) , one can show that @xmath383 . there are two intervals of @xmath384 over which the determinant is greater than zero . one interval is where both the expression in the brackets and the denominator are positive . it is given approximately by the following bound on @xmath385 : @xmath386 the other interval , not mentioned in @xcite , is approximately @xmath387 , where both the denominator and the bracketed expression are negative . the first interval , for @xmath388 , corresponds to rather small values of @xmath258 and @xmath389 using this equation together with eq . ( [ bmu ] ) , one can derive the following result : @xmath390 the above equation is impossible to satisfy in models with the gmsb , because the combination @xmath391 is always negative at the weak scale for the messenger - scale boundary conditions given by eqs . ( [ gauginomass ] ) and ( [ sfermionmass ] ) . to satisfy eq . ( [ pomeq ] ) , a drastic modification of the boundary conditions would be required . we now turn our attention to the second possibility . it requires a relatively large negative value of the singlet soft susy - breaking mass - squared : @xmath392 . this value is impossible to generate unless , as before , one introduces fields @xmath327 and @xmath328 and couples them to @xmath88 . even with the introduction of these fields , if @xmath393 , the extremization conditions can not be simultaneously satisfied . this can be seen in the following way . for @xmath378 ( [ modified_x ] ) takes on the form @xmath394 which implies @xmath395 as @xmath380 . this is incompatible with eq . ( [ mu^2 ] ) , which requires that @xmath396 as @xmath380 . we studied the issue of electroweak symmetry breaking in models with the gauge mediation of supersymmetry breaking ( gmsb ) . we first reviewed various proposals in the literature to generate the @xmath0-parameter of the mssm with the same order of magnitude as the soft susy - breaking parameters such as squark , slepton , and gaugino masses . we find that most of them require small parameters which are accidentally of the same magnitude as the loop factors , cancellation of the kinetic mixing terms at the level of @xmath397 , omission of interactions allowed by symmetries , or many new degrees of freedom not motivated otherwise . even if one could generate the @xmath0-parameter with the same order of magnitude as the soft susy - breaking parameters , it has to have particular values to reproduce @xmath398 gev . we studied this question numerically and found the following . the current experimental lower bounds on superparticle masses limit the overall scale of susy breaking from below , which in turn limits @xmath399 from above ( _ i.e. _ , @xmath400 from below ) . to reproduce @xmath5 , @xmath172 needs to cancel ( too - negative ) @xmath401 and is hence bounded from below . therefore , there is some cancellation required between @xmath172 and @xmath401 . even with the most conservative set of parameters , we found that a cancellation of @xmath402% is necessary . the situation is worse for most of the parameter space this situation was contrasted to the supergravity scenario where the current experimental lower bounds on superparticle masses do not require a significant cancellation among parameters . the simplest mechanism to generate the @xmath0-parameter would be the nmssm , the minimal extension of the mssm without dimensionful parameters in the superpotential . the nmssm is known not to work with the low - energy gmsb , but there was hope that it may work with higher messenger scales . we have shown that this is unfortunately not the case . the current bounds on the superparticles masses are already strong enough to exclude the model completely . we presented a semi - analytic discussion to clarify why the nmssm fails . we also discussed various possible modifications to the nmssm and whether they could lead to a viable electroweak symmetry breaking . the introduction of extra vector - like quarks coupled to the nmssm singlet produces a large negative mass - squared for the singlet , and leads to a viable electroweak symmetry breaking . one needs to adjust the parameters to a few percent , which is comparable to the mssm case for the same @xmath176 range . a fayet illiopoulos @xmath182-term for @xmath217 does not improve the situation . the overall prospect of electroweak symmetry breaking with the gmsb remains unclear . we hope our detailed investigation prompts further studies on this issue . we would like to thank kaustubh agashe and michael graesser for useful discussions . we would also like to thank gian giudice and riccardo rattazzi for helping us understand the mechanism for generating a @xmath0-term proposed in references @xcite and @xcite . this work was supported in part by the director , office of energy research , office of high energy and nuclear physics , division of high energy physics of the u.s . department of energy under contract de - ac03 - 76sf00098 and in part by the national science foundation under grant phy-90 - 21139 . adg was also supported by cnpq ( brazil ) . hm was also supported by the alfred p. sloan foundation . in this appendix we list all of the rg equations for the nmssm , at 1-loop @xcite . these are the equations used , in section 4 , to determine the coupling constants and susy - breaking parameters of the nmssm at the weak scale , given their values at the messenger scale . @xmath403 in the above equations @xmath404 is the @xmath217 gauge coupling ; explicitly @xmath405 . @xmath406 and @xmath407 are , respectively , the weak and strong coupling constants . one defines @xmath408 to be the hypercharge coupling constant in the gut normalization , _ i.e. _ @xmath409 and @xmath410 . gauge couplings at the messenger scale are defined in such a way that they match their experimental values at the @xmath346-mass . we only consider the effect of third generation yukawa couplings , namely , @xmath265 , @xmath411 and @xmath412 . @xmath413 @xmath414 are the soft susy - breaking trilinear couplings , given in sections 2 and 4 . note that we only consider third generation trilinear couplings , namely @xmath415 , @xmath416 , @xmath417 . @xmath418 ( @xmath419=1,2,3 ) are the soft susy - breaking gaugino masses and they evolve , at one loop , identically to @xmath420 . explicitly @xmath421 where @xmath422 is the value of all @xmath423 at the gut scale , while @xmath424 is the common gaugino mass at the gut scale . @xmath426 is the hypercharge - weighted sum of all soft susy - breaking masses - squared @xmath427 where @xmath419 runs over all scalar particles . with the boundary conditions in eqs . ( [ gauginomass],[sfermionmass ] ) , @xmath428 and remains zero throughout the rg evolution . all soft susy - breaking mass - squared terms were taken to be diagonal . again , we only consider the running of third generation soft susy - breaking masses - squared . @xmath429 is defined in section [ sec : nmssm ] . in this appendix we explicitly show the @xmath430 scalar higgs mass - squared matrix of the nmssm . [ 3x3matrix ] @xmath431 where @xmath432 for @xmath433 corresponds , respectively , to @xmath89 , @xmath90 and @xmath91 . all other parameters were defined in previous sections . the determinant of the matrix above can be evaluated explicitly , and its full expression is given bellow . various limits of this determinant are considered in the body of the paper . @xmath434 all parameters were defined previously . recall that @xmath435 . we have studied the nmssm with extra vector - like quarks in section [ extraquarks ] and discussed that the model requires a delicate cancellation among independent parameters . in this appendix , we make further comments on the naturalness of this model . from fig . [ fig:2plots ] , one can easily note that not only does the experimentally allowed value of @xmath332 lie on a steep region of the parameter space , which requires a degree of cancellation of order 1% , but it lies on the _ steepest _ region of the parameter space . one may , therefore , try to address the following question : if all parameters are kept fixed ( and this choice of parameters yields an experimentally allowed spectrum ) except one ( _ e.g. _ @xmath252 ) , what is the likelihood of obtaining a certain value of @xmath332 upon a random choice of the free parameter ? in other words , what is the probability @xmath436 of finding the value of @xmath437 between @xmath332 and @xmath438 given a random choice of @xmath252 ? this line of reasoning is related to the definition of fine - tuning introduced by anderson and castao @xcite . it is easy to note that @xmath439 this `` probability density '' is plotted in fig . [ fig : prob ] . note that we restrict @xmath252 to lie on a range where the same `` qualitative '' physics is obtained , that is , electroweak symmetry is broken and @xmath440 . the plot has been normalized in such a way that @xmath441 . = 0.6 we note that , in some sense , the probability of living in our universe is smaller , if this model is to be taken seriously , than the probability of living in a universe where @xmath442 gev by a factor of three . one can turn this picture around and say that the nmssm , with the above choice of parameters , `` prefers '' ( or predicts ) @xmath442 gev . this does not happen in the mssm . the analog of fig . [ fig:2plots ] would be eq . ( [ eq : mueq ] ) , which is a straight line ( @xmath443 ) if all parameters except @xmath444 are kept fixed . in the language introduced above , the mssm does not `` prefer '' ( or predict ) any particular value of @xmath445 , that is , the `` probability density '' of @xmath445 upon random choices of @xmath444 is flat . in subsection [ extraquarks ] we showed that the values of the higgs boson vevs were extremely sensitive to small variations of the superpotential couplings @xmath252 and @xmath253 . these variations were evaluated numerically after eq . ( [ variation ] ) for one particular set of @xmath252 and @xmath253 . in this appendix we study this issue analytically and show how one can estimate the effects of small variations @xmath446 and @xmath447 on @xmath332 , @xmath176 , and @xmath91 . we will use the following three equations , derived in section [ sec : nmssm ] : @xmath448\ , , \end{aligned}\ ] ] where @xmath449 and we dropped the @xmath93-terms in eq . ( [ d2 ] ) . notice that , because the term on the left - hand side of eq . ( [ d1 ] ) is much smaller then each of the terms on the right - hand side , @xmath450 . for the purpose of the following estimates we will keep only the largest terms in the variations . according to the numbers presented after eq . ( [ variation ] ) , for a small variation of @xmath252 the largest variation on the right - hand side of eq . ( [ d1 ] ) is @xmath451 . we can therefore write @xmath452 we will justify this approximation _ a posteriori_. also , in our analysis we will completely neglect the dependence of the soft - breaking masses - squared on @xmath252 and @xmath253 . this dependence is very weak , as seen in the numbers presented after eq . ( [ variation ] ) . a small change in @xmath252 results in a large change in @xmath332 . hence , to determine the corresponding change in @xmath91 , one can use eq . ( [ d2 ] ) and only consider the variation of @xmath360 , which is approximately given by eq . ( [ dv ] ) . we find @xmath453 so that @xmath454 for the point considered in the text ( @xmath455 ) one finds @xmath456 . next , we consider the effect of @xmath446 on @xmath461 . the problem comes down to estimating @xmath462 , which can be done with the aid of eq . ( [ d3 ] ) : @xmath463 thus , @xmath464 plugging in the numerical values of the parameters , we find that the right - hand side of eq . ( [ d9 ] ) equals @xmath465 . thus , a 1% change in @xmath252 results in a 0.5% change in the value of @xmath461 . since @xmath466 changes by 2% in this case , the contribution of @xmath461 to the variation of @xmath332 is approximately one fourth of that of @xmath466 , consistent with the numbers given in subsection [ extraquarks ] . the above argument can be repeated to find the effect of @xmath447 on @xmath461 . notice that @xmath467 depends on the ratio @xmath468 ( eq . ( [ d3 ] ) ) , and hence changing @xmath253 by @xmath469 has the same effect on @xmath461 as changing @xmath252 by @xmath470 . finally , we show that the condition for @xmath332 to remain constant is @xmath367 . we have already argued that @xmath471 , and therefore @xmath461 , stays unchanged in this case and now show that the same is true for @xmath466 . under @xmath472 the term @xmath473 changes by @xmath474 , while under @xmath475 it changes by @xmath476 . these variations can be made to cancel by imposing @xmath367 . k. inoue , a. kakuto , h. komatsu and s. takeshita , _ prog . * 68 * , 927 ( 1982 ) ; _ prog . phys . _ * 71 * , 413 ( 1984 ) ; + l.e . ibez , _ phys . lett . _ * b118 * , 73 ( 1982 ) ; _ nucl . _ * b218 * , 514 ( 1983 ) ; + l. alvarez - gaum , j. polchinski and m. wise _ nucl . _ * b221 * , 495 ( 1983 ) . k. intriligator , n. seiberg and s. shenker , _ phys . _ * b342 * 152 ( 1995 ) , hep - ph/9410203 ; + h. murayama , _ phys . lett . _ * b355 * , 187 ( 1995 ) , hep - th/9505082 ; + e. poppitz and s. p. trivedi , _ phys . lett . _ * b365 * , 125 ( 1996 ) , hep - th/9507169 ; + k .- i . izawa and t. yanagida , _ prog . _ * 95 * , 829 ( 1996 ) , hep - th/9602180 ; + k. intriligator and s. thomas , _ nucl . phys . _ * b473 * , 121 ( 1996 ) , hep - th/9603158 ; + c. csaki , l. randall , and w. skiba , _ nucl . _ * b479 * , 65 ( 1996 ) , hep - th/9605108 ; + e. poppitz , y. shadmi , and s.p . trivedi , _ nucl . * b480 * , 125 ( 1996 ) , hep - th/9605113 ; + c .- l . chou , _ phys . * b391 * , 329 ( 1997 ) , hep - th/9605119 ; + t. hotta , k .- izawa , and t. yanagida , _ phys . rev . _ * d55 * , 415 ( 1997 ) , hep - ph/9606203 ; + e. poppitz , y. shadmi , and s. p. trivedi , _ phys . lett . _ * b388 * , 561 ( 1996 ) , hep - th/9606184 ; + c. csaki , l. randall , w. skiba , and r. g. leigh , _ phys . lett . _ * b387 * 791 ( 1996 ) , hep - th/9607021 ; + k. intriligator and s. thomas , slac - pub-7143 , hep - th/9608046 ; + e. poppitz and s.p . trivedi , _ phys . * d55 * , 5508 ( 1997 ) , hep - ph/9609529 ; + c. csaki , m. schmaltz , and w. skiba , _ phys . rev . _ * d55 * , 7840 ( 1997 ) , hep - th/9612207 ; + l. randall , _ nucl . * b495 * , 37 ( 1997 ) , hep - ph/9612426 ; + n. haba , n. maru , and t. matsuoka , _ nucl . phys . _ * b497 * , 31 ( 1997 ) , hep - ph/9612468 ; + n. arkani - hamed , j. march - russell , and h. murayama , lbl-39865 , hep - ph/9701286 , to appear in _ nucl . phys . _ * b*. e. poppitz , and s.p . trivedi , _ phys . lett . _ * b401 * , 38 ( 1997 ) , hep - ph/9703246 ; + n. haba , n. maru , and t. matsuoka , _ phys . _ * d56 * , 4207 ( 1997 ) , hep - ph/9703250 ; + y. shadmi , _ phys . _ * b405 * , 99 ( 1997 ) , hep - ph/9703312 ; + r. g. leigh , l. randall , and r. rattazzi , _ nucl . _ , * b501 * , 375 ( 1997 ) , hep - ph/9704246 ; + k .- i . izawa , y. nomura , k. tobe , and t. yanagida , _ phys . _ * d56 * , 2886 ( 1997 ) , hep - ph/9705228 ; + c. csaki , l. randall , and w. skiba , mit - ctp-2654 , hep - ph/9707386 ; + y. nomura , k. tobe , and t. yanagida , ut-797 , hep - ph/9711220 . r. barbieri , s. ferrara , and c. a. savoy , _ phys . lett . _ * 119b * , 343 ( 1982 ) ; + a. h. chamseddine , r. arnowitt , and p. nath , _ phys . lett . _ * 49 * , 970 ( 1982 ) ; + l. j. hall , j. lykken and s. weiberg , _ phys . rev . _ * d27 * , 2359 ( 1983 ) . p. pouliot and n. seiberg , _ phys . lett . _ * b318 * , 169 ( 1993 ) , hep - ph/9308363 ; + d. b. kaplan and m. schmaltz , _ phys . rev . _ * d49 * , 3741 ( 1994 ) , hep - ph/9311281 ; + l. j. hall and h. murayama , _ phys . _ , * 75 * , 3985 ( 1995 ) , hep - ph/9508296 ; + c. d. carone , l. j. hall , and h. murayama , _ phys . rev . _ * d53 * , 6282 ( 1996 ) ; _ ibid_. , * d54 * , 2328 ( 1996 ) , hep - ph/9602364 ; + r. barbieri , g. dvali , and l. j. hall , _ phys . * b377 * , 76 ( 1996 ) , hep - ph/9512388 ; + r. barbieri , l. j. hall , s. raby , and a. romanino , _ nucl . phys . _ * b493 * , 3 ( 1997 ) , hep - ph/9610449 ; + r. barbieri , l. j. hall , a. romanino , _ phys . _ * b401 * , 47 ( 1997 ) , hep - ph/9702315 ; + p. h. frampton and o. c. w. kong , _ phys . rev . _ * d55 * , 5501 ( 1997 ) ; _ phys . * 77 * , 1699 ( 1996 ) , hep - ph/9603372 . n. arkani - hamed , c. d. carone , l. j. hall , and h. murayama , _ phys . _ * d54 * , 7032 ( 1996 ) , hep - ph/9607298 ; + i. dasgupta , b. a. dobrescu , and l. randall , _ nucl . * b483 * , 95 ( 1997 ) , hep - ph/9607487 . p. janot , talk presented at _ hep conference _ , jerusalem ( 1997 ) . + s. assay , talk presented at _ e.p.s . hep conference _ , jerusalem ( 1997 ) . + d. buskulic _ et al . _ , _ phys . * b373 * , 246 ( 1996 ) .

we study the next - to - minimal supersymmetric standard model ( nmssm ) as the simplest candidate solution to the @xmath0-problem in the context of the gauge mediation of supersymmetry breaking ( gmsb ) . we first review various proposals to solve the @xmath0-problem in models with the gmsb . we find none of them entirely satisfactory and point out that many of the scenarios still lack quantitative studies , and motivate the nmssm as the simplest possible solution . we then study the situation in the minimal supersymmetric standard model ( mssm ) with the gmsb and find that an order @xmath1% cancellation is necessary between the @xmath0-parameter and the soft susy - breaking parameters to correctly reproduce @xmath2 . unfortunately , the nmssm does not to give a phenomenologically viable solution to the @xmath0-problem . we present quantitative arguments which apply both for the low - energy and high - energy gmsb and prove that the nmssm does not work for either case . possible modifications to the nmssm are then discussed . the nmssm with additional vector - like quarks works phenomenologically , but requires an order a few percent cancellation among parameters . we point out that this cancellation has the same origin as the cancellation required in the mssm . ucb - pth-97/46 + lbnl-41028 + hep - ph/9711264 + .1 in * next - to - minimal supersymmetric standard model + with the gauge mediation of supersymmetry breaking * 0.3 in andr de gouva , alexander friedland , and hitoshi murayama 0.05 in _ theoretical physics group + ernest orlando lawrence berkeley national laboratory + university of california , berkeley , california 94720 _ 0.05 in and 0.05 in _ department of physics + university of california , berkeley , california 94720 _ .1 in

in the framework of quantum information theory , mixed @xmath1-qubit x - states synthesize a family of quantum states whose inherent correlations are much easier to quantify than is generally the case . the prefix ` x ' is motivated by the shape of their density matrices written in the computational basis @xcite , whose non - zero entries are either diagonal or anti - diagonal ( or , otherwise , can be brought to this form via a local unitary ( lu ) transformation ) . owing to this sparse structure , that includes important states ( e.g. , bell s @xcite , werner s @xcite , isotropic @xcite , ghz @xcite , etc . ) , _ analytical _ investigations of entanglement properties@xcite and quantum discord @xcite in @xmath1-qubit x - states have lately become an active and fruitful field of research . retrospectively , important members of the class of two - qubit x - states were identified in @xcite , where the concept of maximally entangled mixed states ( mems ) was introduced and characterized . from these seminal works , worthy of note is the observation that two - qubit states of a fixed spectrum and maximal entanglement ( as measured by concurrence , negativity or relative entropy of entanglement ) can always be found in the x - form . subsequently , munro _ et al . _ @xcite characterized two - qubit states of maximal entanglement for a fixed mixedness ( as measured by purity , linear entropy or von neumman entropy ) , once again obtaining x - states as results . in spite of these early achievements , to date , little has been accomplished in extending the characterization of mems beyond two - qubits . largely , this is because sensible measures of _ genuine multipartite _ entanglement have been identified only recently @xcite and are generally hard to evaluate , let alone maximize . a first important step toward the identification of @xmath1-qubit mems for @xmath2 was given by hashemi rafsanjani _ et al_. @xcite , who showed that the gm - concurrence of @xmath1-qubit _ x - states _ admits a simple closed formula , amenable to maximization . although the resulting optimal states of this maximization can not be guaranteed to be actual mems , at least they are provably mems amongst all @xmath1-qubit x - states . therefore , in @xcite , agarwal and hashemi rafsanjani maximized the x - state gm - concurrence formula under the constraint of a fixed linear entropy , determining the so - called x - mems . in this paper , we enlarge the scope of the term x - mems to enclose two classes of x - states : _ x - mems with respect to _ ( wrt ) _ spectrum _ , referring to those @xmath1-qubit x - states of maximal gm - concurrence for a fixed spectrum , in analogy to the original mems introduced in @xcite ; and _ x - mems wrt purity _ , referring to @xmath1-qubit x - states that maximize the gm - concurrence for a fixed value of purity , in parallel with @xcite . our main results initially consist of : ( i ) a complete characterization of x - mems wrt spectrum , and ( ii ) a demonstration that x - mems wrt purity can be obtained from the solution of a semidefinite program ( sdp ) @xcite , by which means ( iii ) we provide an alternative proof of optimality of the states obtained in @xcite . moreover , ( iv ) we characterize the unitary transformation that maximizes the x - state gm - concurrence formula of an arbitrary @xmath1-qubit state , generalizing the result of @xcite for @xmath3 . finally , of independent interest ( but also relevant in this context ) , ( v ) we construct a family of iterated sdps whose solutions produce quantum operations ( cptp maps ) that implement a desired state transformation with a decreasing number of kraus operators . the method is illustrated with the determination of short operator - sum representations for the conversion between an arbitrary input state of purity @xmath4 and a corresponding x - mems wrt purity @xmath4 . our paper is structured as follows . in section [ sec : preliminaries ] , we briefly review the concept of gm - concurrence and , in particular , its simple formula for @xmath1-qubit x - states . in section [ sec : xmemswrtspectrum ] , we characterize x - mems wrt spectrum and the unitary transformations that produce such states from arbitrary @xmath1-qubit density matrices . in section [ sec : xmemswrtpurity ] , x - mems wrt purity are constructed and have their optimality established via sdp theory . section [ sec : convert ] outlines the iterated sdp method to design quantum state transformation with few kraus operators and exemplifies the method while producing x - mems wrt purity from arbitrary input states of the same purity . finally , section [ sec : conclusion ] summarizes our results and discusses some possible avenues of future work . in this section , we present the formula for the gm - concurrence of @xmath1-qubit x - states obtained in @xcite . for the benefit of the reader unfamiliar with the current literature on multipartite entanglement ( in particular , @xcite ) , we start by reviewing some key definitions concerning @xmath1-qubit x - states , gm - entanglement and gm - concurrence . to begin with , let us introduce some notation . throughout , @xmath5 denotes the ( complex ) hilbert space of dimension @xmath6 , whereas @xmath7 denotes the set of ( bounded ) linear operators acting on @xmath5 . the set of all possible bipartitions of @xmath8 is denoted by @xmath9 and a particular bipartition @xmath10 in @xmath9 is denoted by @xmath11 ( with @xmath12 ranging from @xmath13 to @xmath14 ) . partial traces over hilbert spaces @xmath5 whose labels @xmath15 belong to @xmath16 are concisely indicated as @xmath17 . an operator @xmath18 represents an _ @xmath1-qubit x - state _ if and only if , in the computational basis @xmath19 ( and up to a lu - transformation ) , it assumes the matrix form @xmath20\,,\ ] ] where @xmath21 and , for every integer @xmath22 $ ] , we have @xmath23 , @xmath24 , @xmath25 , @xmath26 $ ] , @xmath27 while ( [ eq : nqubitxstate ] ) visually justifies the prefix ` x ' , the conditions ( [ eq : normandpsd ] ) ensure the normalization and positive semidefiniteness of @xmath28 . as a glance at ( [ eq : nqubitxstate ] ) demonstrates , the index @xmath22 $ ] can be regarded as a label for uncoupled bidimensional subspaces . that any @xmath1-qubit x - state is decomposable into @xmath29 such subspaces is a key property that will be implicitly exploited throughout this paper . although we are only interested in the entanglement properties of @xmath1-qubit x - states , we proceed with a general definition of gm - entanglement . [ def : gment ] an @xmath1-partite density operator @xmath30 is _ gm - entangled _ if and only if it is not _ biseparable_. to understand the concept of biseparability , consider first its definition for pure states . [ def : biseppure ] an @xmath1-partite state @xmath31 is _ biseparable _ if and only if there is a hilbert space bipartition @xmath32 and a pair of states @xmath33 , @xmath34 , such that @xmath35 . note that definition [ def : biseppure ] implies that a biseparable state is not necessarily separable , as there might be entanglement within @xmath36 and/or @xmath37 . it then follows from definition [ def : gment ] , that the condition for gm - entanglement is generally more stringent than the condition for bipartite entanglement , for example . in fact , gm - entanglement only occurs when bipartite entanglement is observed across all possible bipartitions of @xmath38 . the notion of biseparability is extended to mixed states as follows . an @xmath1-partite density operator @xmath39 is _ biseparable _ if and only if it can be decomposed in an ensemble of biseparable pure states , that is @xmath40 where @xmath41 and each @xmath42 is biseparable ( even if with respect to different bipartitions of @xmath38 ) . the above definitions provide a formal criterion to determine whether a general mixed state is gm - entangled or not . a further step was given by ma _ et al . _ @xcite , who introduced the gm - entanglement measure named gm - concurrence . the gm - concurrence of an @xmath1-partite pure state @xmath31 is given by @xmath43}\,,\ ] ] where @xmath44 $ ] . for @xmath1-partite density operators @xmath39 , the gm - concurrence is obtained via the convex roof construction @xmath45 with the infimum taken over all possible ensembles @xmath46 that realize @xmath47 . the gm - concurrence takes its name from the fact that , in the case of two - qubit systems , it matches the wootters concurrence @xcite and , more generally , can be shown @xcite to satisfy the following minimal requirements for any gm - entanglement measure : * gm - entanglement detection : @xmath48 with saturation if and only if @xmath47 is biseparable . * convexity : @xmath49 * monotonicity under local operations and classical communication ( @xmath50 ) : @xmath51)\leq c_{\rm gm}(\bm{\rho})\,.\ ] ] * invariance under lu - transformations ( @xmath52 ) : @xmath53 though well motivated , @xmath54 is generally hard to evaluate due to the infimum over all ensembles that realize @xmath47 . to alleviate this problem , the authors of @xcite relied on certain sufficient criteria for gm - entanglement detection proposed by huber _ et al . _ @xcite to determine computable lower bounds for @xmath55 . in particular , if the main- and anti - diagonal entries of @xmath47 are parametrized as in ( [ eq : nqubitxstate ] ) ( the remaining entries being arbitrary ) , then one of ma s lower bounds reads ( see ( * ? ? ? * appendix a ) for an explicit derivation ) @xmath56}\left[r_k-\sum_{j\neq k}^n\sqrt{a_jb_j}\right]\right\}\,.\ ] ] remarkably , as shown by hashemi rafsanjani _ et al . _ @xcite , this lower bound is saturated when @xmath57 , namely , @xmath58}\left[r_k-\sum_{j\neq k}^n\sqrt{a_jb_j}\right]\right\}\,.\ ] ] the fact that @xmath1-qubit x - states have their gm - concurrence expressed as a closed formula can not be overstated . it contrasts with the great difficulty involved in merely detecting gm - entanglement in more general systems , not to mention quantifying it . of course , this result becomes even more appealing when one notices that @xmath1-qubit states of practical interest do occur in the x - form ( see , e.g. , @xcite ) , or otherwise can usually be well approximated to it via lu - transformations @xcite . finally , it is interesting that for ghz - diagonal states ( x - states with @xmath59 ) , the value of gm - concurrence is proportional to the distance of the ghz - state to the set of biseparable states @xcite . as mentioned before , the two - qubit mems with a given spectrum , characterized in @xcite , are x - states . in this section , we _ assume _ that this is also true in the @xmath1-qubit case ( @xmath2 ) , and characterize the `` @xmath1-qubit mems '' resulting from this assumption . since it is not known in which circumstances the restriction to the set of x - states is an active constraint for @xmath2 , we adopt the nomenclature introduced in @xcite and talk about x - mems instead of simply mems . the results of this section are summarized in theorem [ thm : xmemswrtspectrum ] , which is deliberately presented in close resemblance to the statement of the related theorem presented in @xcite , regarding the case @xmath3 . the proofs , however , are established in very different ways . [ thm : xmemswrtspectrum ] the maximal gm - concurrence attainable by an @xmath1-qubit x - state of spectrum @xmath60 , determined by the eigenvalues @xmath61 ( with @xmath21 ) , is given by @xmath62\,.\ ] ] moreover , any @xmath1-qubit density matrix @xmath63 ( of spectrum @xmath60 ) can be coherently transformed into @xmath64 , an isospectral @xmath1-qubit x - density matrix of maximal gm - concurrence , according to @xmath65 , with the unitary @xmath66 given by @xmath67 \bm{d}_\phi \bm{\phi}^\dagger\,.\ ] ] in ( [ eq : globalunitary ] ) , the following definitions apply : @xmath68 is a set of arbitrary single qubit unitary operations , @xmath69 is an arbitrary unitary and diagonal matrix , @xmath70 is the unitary matrix formed from the eigenvectors of @xmath63 ( such that @xmath71 ) , and @xmath72 here , @xmath73 is the @xmath29-dimensional matrix whose only non - zero entry is equal to @xmath13 and occupies the @xmath15th row and @xmath74th column . an immediate remark is that , as expected , both the optimal gm - concurrence ( [ eq : cgmoptimal ] ) and the unitary transformation ( [ eq : globalunitary ] ) reduce to the corresponding expressions in ( * ? ? ? * theorem 1 ) when @xmath3 . the remainder of this section is devoted to proving the theorem for @xmath2 . essentially , our proof consists of a direct maximization of the gm - concurrence formula of @xmath1-qubit x - states under the constraint of a fixed spectrum . let @xmath75 denote a generic @xmath1-qubit x - density matrix of spectrum @xmath76 . we start by taking matrix ( [ eq : nqubitxstate ] ) as a parametrization for @xmath75 and writing a general formula for @xmath77 in terms of its eigenvalues @xmath78\,.\ ] ] here , @xmath79 denote the greatest ( @xmath80 ) and smallest ( @xmath81 ) eigenvalues of @xmath75 associated with the bidimensional subspace labelled by @xmath82 , and @xmath83 . it follows trivially from ( [ eq : eigslambdak ] ) that , @xmath84 which used in ( [ eq : cgenqubitxstate ] ) yields @xmath85,\ ] ] where , without loss of generality , we fixed the label @xmath86 to the subspace whose value of @xmath87 is the largest . our goal now is to maximize ( [ eq : eigscgenqubitxstate ] ) under the contraint that the set of eigenvalues @xmath88 matches the set of eigenvalues of the arbitrary ( but given ) @xmath1-qubit density matrix @xmath63 , i.e. , @xmath89 . to do that , let us first consider the optimization over @xmath90 and @xmath91 . although these variables are constrained by ( [ eq : normandpsd ] ) and related to @xmath88 by ( [ eq : eigslambdak ] ) , we will momentarily ignore these contraints . by doing so , we significantly simplify the optimization procedure at the expense of risking over - maximization @xmath55 . nevertheless , as we will soon demonstrate , the resulting maximal is actually attainable , meaning that our simplifying assumptions are harmless . with that in mind we set , for every @xmath92 $ ] , @xmath93 which clearly maximizes ( [ eq : eigscgenqubitxstate ] ) over @xmath94 and @xmath90 . at this point , we are left with the maximization over @xmath88 , written as @xmath95 where @xmath96 are the eigenvalues of the arbitrary ( but given ) @xmath1-qubit density matrix @xmath63 and the vectors @xmath97 are given by @xmath98 here , we aim to assign to each variable in @xmath88 an eigenvalue of @xmath63 , in such a way that @xmath99 is maximal and @xmath100 is minimal . to maximize @xmath99 , we simply assign to it the largest eigenvalue of @xmath63 , i.e. , @xmath101 to minimize @xmath100 , first notice that @xmath102 and @xmath103 display the same entries in the reversed order , with @xmath104 occupying the central position in both vectors . it follows from the rearrangement inequality ( see , e.g. , ( * ? ? ? * theorem 368 , page 261 ) ) that the scalar product between two vectors defined up to the ordering of their entries is minimized if and only if they are sorted in opposite directions . so , we make the entries of @xmath102 and @xmath103 monotonically increasing and decreasing , respectively , by assigning , for every @xmath92 $ ] , @xmath105 substituting the identities ( [ eq : rjd1 ] ) , ( [ eq : lambda1peqlambda1 ] ) and ( [ eq : ansatzdistrib ] ) in ( [ eq : mixabreigs ] ) and solving the resulting system for @xmath106 , @xmath23 and @xmath24 ( under the constraints described in ( [ eq : normandpsd ] ) ) , we obtain that and @xmath107 indicated in ( [ eq : solsystem ] ) for any @xmath92 $ ] . however , this does not lead to essentially new x - mems , since the x - mems corresponding to these solutions can always be generated from the x - mems corresponding to ( [ eq : solsystem ] ) via a lu - transformation . ] , for every @xmath92 $ ] , @xmath108 by plugging ( [ eq : solsystem ] ) into ( [ eq : cgenqubitxstate ] ) , it is easily seen that ( [ eq : cgmoptimal ] ) holds . in order to see that ( [ eq : cgmoptimal ] ) is physically attainable , substitute ( [ eq : solsystem ] ) into matrix ( [ eq : nqubitxstate ] ) ( and set @xmath109 for every @xmath22 $ ] ) , to get @xmath110\,.\ ] ] it is immediate to check that @xmath64 is a valid x - density matrix with the same spectrum of @xmath63 and gm - concurrence given by ( [ eq : cgmoptimal ] ) . finally , let us establish ( [ eq : globalunitary ] ) . since @xmath63 and @xmath64 are isospectral , we can write @xmath111 where @xmath70 is the matrix of the eigenvectors of @xmath63 and @xmath112 is the matrix of the eigenvectors of @xmath64 given by ( [ eq : rhoxprime ] ) . some simple linear algebra shows that , for @xmath113 $ ] , the matrix @xmath112 admits the block decomposition specified in ( [ eq : v ] ) . thus , combining the two identities in ( [ eq : rhoandrhoprime ] ) to eliminate @xmath60 , we arrive at @xmath114 \bm{\phi}^\dagger\,.\ ] ] as noted before , the x - mems of ( [ eq : rhoxprime ] ) are unique up to lu - transformations , for which reason , in ( [ eq : globalunitary ] ) , the expression of @xmath66 appears pre - multiplied by an arbitrary lu - transformation . furthermore , for sake of generality , we have also multiplied an arbitrary ( generally non - local ) diagonal unitary matrix @xmath69 in ( [ eq : globalunitary ] ) . it should be emphasized , though , that @xmath69 has obviously no effect on the output state @xmath64 . let us conclude this section by answering a central question that arises from the present work : are @xmath1-qubit x - mems actual @xmath1-qubit mems ? although this is long known to be the case for @xmath3 @xcite , indications that the same may also hold for @xmath115 have only recently appeared in the work of hedemann @xcite . alas , to the best of our knowledge , the topic seems to be utterly unexplored for @xmath116 . to see that that conjecture can not hold in general , note that an affirmative answer ( combined with ( [ eq : cgmoptimal ] ) ) , would imply that @xmath1-qubit density matrices whose eigenvalues @xmath61 satisfy @xmath117 can not acquire gm - entanglement by means of a global unitary transformation . however , as recently shown by huber _ et al . _ @xcite , @xmath1-qubit thermal states of arbitrarily high temperatures and @xmath1 sufficiently large ( represented , in the computational basis , by diagonal density matrices arbitrarily close to the identity , hence fulfilling ( [ eq : sfs ] ) ) _ can _ acquire gm - entanglement by means of rotations to dicke - like ( non - x ) states , thus providing a counter - example to the original conjecture . since the purity of a given quantum state can be expressed as the sum of the squares of their eigenvalues , there are , in general , many spectra that realize a fixed value of purity @xmath4 . in this section , we consider the problem of searching for a maximizer of the @xmath1-qubit x - state gm - concurrence formula amongst all @xmath1-qubit spectra that realize @xmath4 . any x - state with such an optimal spectrum is referred to as an x - mems wrt purity , denoted by @xmath118 . as a matter of fact , @xmath1-qubit x - mems wrt purity have been recently determined in the work of agarwal and hashemi rafsanjani @xcite , whose main findings are summarized in the statement of the following theorem . [ thm : xmemswrtpurity ] for any value of purity @xmath1191/(n+1),1]$ ] with @xmath21 , the maximal gm - concurrence attainable by an @xmath1-qubit x - state of purity @xmath4 is @xmath120 , where the parameter @xmath1210,1/2]$ ] is determined by @xmath4 according to @xmath122 \frac{1}{2n}+\frac{1}{2}\sqrt{(1-\frac{1}{n})(p-\frac{1}{n})}&\mbox{if}\quad \frac{n+3}{(n+1)^2}\leq p\leq 1\\ \end{array } \right.\,.\ ] ] up to lu - transformations , every @xmath1-qubit x - density matrix of purity @xmath4 that achieves maximal gm - concurrence is given by ( [ eq : rhoxprime ] ) with @xmath123 for every @xmath92 $ ] , being @xmath124 and @xmath125 defined as follows : @xmath126 \gamma&\mbox{if}\quad \frac{1}{n+1}\leq \gamma\leq \frac{1}{2}\\ \end{array } \right.\quad\mbox{and}\quad g(\gamma)\coloneq\frac{1 - 2f(\gamma)}{n-1}\,.\ ] ] in what follows , we give an alternative proof of this theorem by exploiting the results of theorem [ thm : xmemswrtspectrum ] and the theory of semidefinite programming @xcite . since purity is determined by the spectrum ( and not the other way around ) , the specification of a spectrum generally represents a stronger constraint than a specification of purity . as a result , if @xmath118 is a _ x - mems wrt purity @xmath4 _ and has spectrum @xmath127 , then @xmath118 is also a _ x - mems wrt to the spectrum @xmath127_. in other words , every x - mems wrt purity can be regarded as a x - mems wrt _ some _ spectrum , in which case every @xmath118 is of the form ( [ eq : rhoxprime ] ) up to a lu - transformation . thanks to this , we can determine @xmath118 by maximizing the gm - concurrence of ( [ eq : rhoxprime ] ) over all sets of physical eigenvalues that realize @xmath4 , which yields the optimization problem represents the set of optimization variables , as opposed to a fixed set of eigenvalues ( as it was the case in sec . [ sec : xmemswrtspectrum ] ) . ] @xmath128 next , we give a few arguments that allow some simplification of problem ( [ eq : optprblm2 ] ) . first , there is no need to explicitly require the ordering @xmath129 since we know in advance ( rearrangement inequality ) that this particular ordering for any @xmath92 $ ] . ] is a necessary condition for the maximization of the considered objective function and will be thus satisfied anyway . second , the objective function is clearly maximized if we set @xmath130 for every @xmath92 $ ] , which does not violate any problem constraint and reduces the equality constraints to @xmath131 given these two points , we can replace the original ( non - linear ) objective function with the linear function @xmath132 , and all the inequality constraints with a single one : @xmath133 . finally , without altering the solution of the problem , we can replace the equality in the quadratic constraint with the inequality ` @xmath134 ' , thus enlarging the set of feasible points to its convex hull ( * ? ? ? * chapter 32 ) . accordingly , we end up with the equivalent optimization problem on @xmath29 real variables : @xmath135 as explained in [ app : quadtolmi ] , the quadratic ( convex ) constraint in ( [ eq : optprblm3 ] ) can be written as a linear matrix inequality ( lmi ) , turning problem ( [ eq : optprblm3 ] ) into the sdp ( [ eq : sdpopennotation ] ) . it is straightforward to see that ( [ eq : sdpopennotation ] ) admits the _ _ sdp standard inequality form _ _ ) and ( [ eq : optprblm4 ] ) are solved by the same set of eigenvalues , the resulting optimal values of the two problems are not exactly the same . that is because , to arrive at problem ( [ eq : optprblm4 ] ) , we removed the summand @xmath136 from the objective function of ( [ eq : sdpopennotation ] ) ( and used the property @xmath137 to replace the maximization with the minimization ) . as a result , if @xmath138 and @xmath139 denote the optimal value of problems ( [ eq : sdpopennotation ] ) and ( [ eq : optprblm4 ] ) , respectively , then @xmath140.[foot : objfxn ] ] @xmath141 with @xmath142\quad\mbox{and}\quad \bm{f}_i\coloneq \left[\begin{array}{c|c } \begin{array}{c|c } \bm{0}_{n}&\bm{e}_{n , i}\\\hline \bm{e}_{n , i}^{\sf t}&2 \end{array } & \\\hline & -1 \end{array}\right]\,,\ ] ] where @xmath143 , @xmath144 , @xmath145 were defined in the [ app : quadtolmi ] and @xmath146 denotes the @xmath29-dimensional vector whose only non - zero entry is @xmath13 and occupies the @xmath15th row . the fact that the design of the maximally gm - entangled x - states of @xmath1-qubits can be cast as a sdp is interesting on its own . first , sdps are convex optimization problems @xcite and , as such , have the desirable property that any local optimum is necessarilly a global optimum . second , many efficient numerical methods have been devised to solve sdps ( for a review , see , e.g. , @xcite and references therein ) . these methods have excellent convergence properties and output a `` certificate of convergence '' , i.e. , an interval within which the optimal value of the objective function must lie . typically , with no more than @xmath147 iterations , this interval can be made arbitrarily small . third ( and most importantly for our purposes ) , a powerful duality theory exists for sdps and can be employed to rigorously prove the optimality of an _ ansatz _ solution . before proceeding with our proof , let us exploit the aforementioned numerical virtues of sdps to provide a first evidence of the optimality of the gm - concurrence @xmath120 ( cf . ( [ eq : gammafxnp ] ) ) and of the spectrum ( [ eq : propprimalsol ] ) . in figure [ fig : evidopt ] , we plot these analytical expressions ( lines ) along with the numerical solutions ( symbols ) of problem ( [ eq : optprblm4 ] ) , obtained by running the matlab - based solver sedumi @xcite for several combinations of @xmath4 and @xmath1 . in our numerical computations , we set sedumi s precision to @xmath148 , which establishes the largest acceptable length of the aforementioned `` error interval '' . ) and the corresponding analytical formulas ( lines ) for the optimal gm - concurrence and spectrum ( cf . theorem [ thm : xmemswrtpurity ] ) . each plot considers the parameter values @xmath149 and a uniform sampling of @xmath1501/(n+1),1]$ ] . the legend in plot ( a ) also applies to plots ( b ) and ( c ) . the observed agreement between the lines and symbols strongly suggests the optimality of the analytical formulas in the statement of theorem [ thm : xmemswrtpurity ] . from left to right , the unlabelled tick marks correspond to the purity value @xmath151 for @xmath152 ( or , equivalently , @xmath153 ) . remarkably , although all the plotted functions are continuous , they fail to be smooth : the second derivative of @xmath154 and the first derivatives of @xmath155 are discontinuous at @xmath156 . ] plot [ fig : evidopt]a shows the converged numerical values of the objective function along with an analytic plot of @xmath157 . the difference between the numerical and analytical values is found to be of the order of @xmath158 , which strongly suggests that @xmath157 is the maximal @xmath1-qubit x - state gm - concurrence wrt purity . plots [ fig : evidopt]b , c , in turn , indicate the optimality of ( [ eq : propprimalsol ] ) . in particular , plot [ fig : evidopt]b reveals an excellent agreement between the converged numerical values of @xmath159 and the function @xmath160 . similarly , plot [ fig : evidopt]c illustrates the coincidence between the converged numerical values of @xmath161 ( which resulted all the same up to numerical precision ) and the function @xmath162 , cf . ( [ eq : propprimalsol ] ) and ( [ eq : defsfandg ] ) . we now briefly review an important result from the sdp duality theory that will be subsequently used to establish the optimality of ( [ eq : propprimalsol ] ) . for a thorough account on this theory , we refer the reader to ( * ? ? ? * chapter 5 ) . the _ dual problem _ of the sdp ( [ eq : optprblm4 ] ) ( henceforth called _ primal problem _ ) , is another sdp given by @xmath163\quad\mbox{subject to}\quad \bm{z}\geq 0\quad\mbox{and}\quad \tr[{\bm{f}_i\bm{z}}]=c_i\quad \forall i=1,\ldots , n\,.\ ] ] here , the variable to be optimized is the matrix @xmath164 , whereas the vector @xmath165 and the matrices @xmath166 and @xmath167 are the same as the ones appearing in the primal problem ( in our particular case , defined in ( [ eq : sdpgivens ] ) ) . let @xmath168 denote any feasible value of the primal problem ( [ eq : optprblm4 ] ) , and denote by @xmath139 its optimal value . similarly , let @xmath169 and @xmath170 denote , respectively , any feasible value and the optimal value of the dual problem ( [ eq : dualprblm ] ) . it is obvious that @xmath171 and that @xmath172 . less obvious but also true=\sum_{i=1}^n\tr\left[\bm{f}_i \bm{z}\right]\lambda_i+\tr\left[\bm{f}_0\bm{z}\right]=\tr\left[\left(\bm{f}_0+\sum_{i=1}^n\bm{f}_i\lambda_i\right)z\right]\geq 0\,.\ ] ] ] is that @xmath173 for _ every _ primal and dual feasible values ( in particular , @xmath174 ) , in such a way that @xmath175 a general result known as _ weak duality_. next , we show how these inequalities can be used to prove the optimality of our _ ansatz _ solution . as some straightforward computation shows , the spectrum ( [ eq : propprimalsol ] ) can be initially regarded as a _ primal feasible point _ that yields a _ primal feasible value _ @xmath176 . however , if we can find a particular _ dual feasible point _ that yields a _ dual feasible value _ @xmath177 , then weak duality implies that @xmath178 , meaning that ( [ eq : propprimalsol ] ) is , indeed , a _ primal optimal point_. we claim that the block matrix @xmath179 , described below , provides such dual feasible point : @xmath180\,,\ ] ] where @xmath181 and @xmath182 denote the `` all - one '' @xmath29-dimensional vector and matrix , respectively . in addition , @xmath183 \frac{(1-n)(1 + 2\gamma)^2}{2(1 - 2n\gamma)}&\mbox{if}\quad \frac{1}{n+1}\leq \gamma\leq \frac{1}{2 } \end{array}\right.\ , , & \quad \mathfrak{z}_{3}\coloneq\left\{\begin{array}{ll } \frac{1}{2\gamma}&\mbox{if}\quad 0 < \gamma\leq \frac{1}{n+1}\\[2 mm ] \frac{1-n}{2(1 - 2n\gamma)}&\mbox{if}\quad \frac{1}{n+1}\leq \gamma\leq \frac{1}{2 } \end{array}\right . , \\[10 mm ] \fl \mathfrak{z}_{2}\coloneq\left\{\begin{array}{ll } \frac{(1+\gamma)^2}{2\gamma}&\mbox{if}\quad 0 < \gamma\leq \frac{1}{n+1}\\[2 mm ] \frac{(n-2\gamma)^2}{2(1-n)(1 - 2n\gamma)}&\mbox{if}\quad \frac{1}{n+1}\leq \gamma\leq \frac{1}{2 } \end{array}\right.\ , , & \quad \mathfrak{z}_{4}\coloneq \left\{\begin{array}{ll } 0&\mbox{if}\quad 0 < \gamma\leq \frac{1}{n+1}\\[2 mm ] 1+\frac{1 - 2\gamma}{1 - 2n\gamma}&\mbox{if}\quad \frac{1}{n+1}\leq \gamma\leq \frac{1}{2 } \end{array}\right.\ , . \end{array}\ ] ] to verify our claim , we first show that @xmath179 establishes @xmath177 . indeed , @xmath184 $ ] , which can be evaluated with the aid of ( [ eq : sdpgivens ] ) and ( [ eq : matrixz ] ) to give @xmath185-\mathfrak{z}_3(p-1)-\mathfrak{z}_4\,.\label{eq : dualvaluez}\ ] ] then , plugging ( [ eq : z ] ) into ( [ eq : dualvaluez ] ) and expressing @xmath4 as the sum of the squares of the eigenvalues in ( [ eq : propprimalsol ] ) , we obtain @xmath186 , as claimed . now , we show that @xmath179 satisfies the constraints of problem ( [ eq : dualprblm ] ) . regardless of the values @xmath187 , it is easy to check that @xmath188=-2 $ ] and @xmath189=-1 $ ] , as required . finally , the condition @xmath190 can be checked by an explicit study of the eigenvalues of @xmath179 . simple inspection of ( [ eq : z ] ) shows that @xmath191 . furthermore , with some cumbersome simplification procedure , the characteristic polynomial of the first @xmath192-dimensional block of @xmath179 takes the form @xmath193 with @xmath194 \frac{n^2 + 2(n-1)-4\gamma+4n\gamma^2}{2(-1 + 2n\gamma)}&\mbox{if}\quad \frac{1}{n+1}\leq \gamma\leq \frac{1}{2 } \end{array}\right.\,,\ ] ] thus establishing @xmath195 as the only non - zero eigenvalue of this block . clearly , the first branch of ( [ eq : lambdan ] ) is strictly positive . to see that this is also true for the second branch , define @xmath196 in such a way that , for @xmath197 $ ] , @xmath198 . we conclude our proof by showing that , for @xmath199 , both @xmath200 and @xmath201 are strictly positive , hence so is @xmath202 . to establish the positivity of @xmath201 , note the following implications : @xmath203 the first implication follows from the fact that , for the relevant values of @xmath29 ( recall that @xmath204 ) , the inequality @xmath205 holds trivially . the positivity of @xmath200 , in turn , can be established by noting that @xmath206}{(n+1)^2}>0\quad\mbox{and}\quad \frac{\partial f_n(\gamma)}{\partial \gamma } = 2 g_n(\gamma)>0\,.\ ] ] so , we see that @xmath200 is already positive at @xmath207 and monotonically increasing for @xmath208 . for any given pair of isospectral density matrices , it is possible to find a unitary transformation that maps one density matrix into the other . in theorem [ thm : xmemswrtspectrum ] , for example , all unitary transformations that map an arbitrary @xmath1-qubit density matrix into a corresponding x - mems wrt spectrum were explicitly constructed . however , if two density matrices have different spectra ( e.g. , an arbitrary density matrix and a corresponding x - mems wrt purity , cf . [ sec : xmemswrtpurity ] ) , then there is no unitary map capable of converting between the two , in which case one must resort to more general quantum operations to implement the desired state transformation . in this section , we introduce a numerical scheme to design a quantum channel @xmath209 , as modelled by a completely positive and trace preserving ( cptp ) map , that promotes the conversion between any two given density matrices . furthermore , our scheme constrains the resulting map to be `` economical '' wrt certain resources utilized in its implementation . to understand what _ economical _ means in this context , consider the general representation of @xmath209 in terms of kraus operators @xmath210 : @xmath211 a trivial choice of @xmath209 , such that @xmath212 , is the channel that maps _ every _ density matrix to @xmath213 , which admits the following _ minimal _ set of kraus operators : @xmath214 where @xmath215 , @xmath216 , @xmath217 is an orthonormal basis formed from the eigenvectors of @xmath213 and @xmath218 are the non - zero eigenvalues of @xmath213 ( corresponding to the eigenvectors @xmath219 ) , i.e. , @xmath220 practically , though , implementing ( [ eq : sillymm ] ) can be considered an overkill . in fact , since we _ only _ require @xmath221 , it might be possible to find a quantum channel that implements the desired state transformation with a number of kraus operators much smaller than @xmath222 . in other words , there might be a more _ economical cptp map _ for the state transformation @xmath221 . with that mind - set , consider the task of determining , amongst every cptp map @xmath209 that satisfies @xmath223 , those that can be decomposed with the smallest possible number of kraus operators . mathematically , this leads to an optimization problem that can be nicely expressed with the aid of the choi - jamiokoswki isomorphism @xcite , which brings cptp maps @xmath224 into a one - to - one correspondence with positive semidefinite matrices @xmath225 such that @xmath226={\mathbf{1}}_d$ ] ( here and throughout , @xmath227 denotes the partial trace over the @xmath228th @xmath169-dimensional subsystem ) . in particular , given a cptp map @xmath224 , its corresponding choi - jamiokoswki matrix is @xmath229 where @xmath230 is the identity map on @xmath231 and @xmath232 is the ( unnormalized ) maximally entangled state @xmath233 , with @xmath234 a fixed orthonormal basis for @xmath235 . in this framework , the minimal kraus decompositions of @xmath209 can be shown to have @xmath236 kraus operators @xcite , and the equation @xmath223 is equivalent to @xmath237=\overline{\bm{\rho}}$ ] ( the transposition being taken wrt @xmath47 written in the basis @xmath234 ) @xcite . accordingly , the optimization problem takes the form @xmath238={\mathbf{1}}_d\quad\mbox{and}\quad \tr_1\left[(\bm{\rho}^{\sf t}\otimes { \mathbf{1}}_d)\bm{\mathfrak{c}}\right]=\overline{\bm{\rho}}\,.\label{eq : rnkproblem}\end{aligned}\ ] ] problem ( [ eq : rnkproblem ] ) is an example of rank minimization problem ( rmp ) with sdp constraints . although special cases of this problem have been solved ( see , e.g. , @xcite and references therein ) , rmp are , in general , computationally intractable ( np - hard ) due to the non - smoothness and non - convexity of the rank function . fortunately , though , heuristic methods exist to efficiently approximate their solutions @xcite . essentially , these heuristics rely on replacing the rank with a surrogate function , in such a way that the resulting problem can be handled with standard sdp solvers . next , we consider the application of two such methods to problem ( [ eq : rnkproblem ] ) . basically , it consists of replacing @xmath236 with @xmath239 $ ] . intuitively , the replacement makes sense because sparse vectors tend to have a @xmath240-norm smaller than dense vectors . so , by forming a vector from the eigenvalues of @xmath241 and minimizing its @xmath240-norm ( which corresponds to minimize @xmath239 $ ] when @xmath242 ) , we may be effectively vanishing some eigenvalues of @xmath241 and , hence , reducing its rank ( see @xcite for a more technical justification of the trace heuristic in terms of the convex envelope of the rank ) . to a large extent , the appeal of the trace heuristic comes from the fact that it provides a _ linear _ objective function for the optimization problem , which usually means that the solution can be efficiently obtained ( at least numerically ) . unfortunately , in the case of problem ( [ eq : rnkproblem ] ) , the constraint @xmath226={\mathbf{1}}_d$ ] implies that @xmath239=d$ ] , so no minimization can actually occur . as a result , the trace heuristic is useless for our purposes . introduced in @xcite , the log - det heuristic can be considered a refinement of the trace heuristic . consists of replacing @xmath236 with @xmath243 , where @xmath244 is a regularization constant . the value of @xmath245 can be made arbitrarily small , and it is used to avoid an ill - defined objective function as @xmath241 gets singular ( @xmath246 ) . intuitively , it is expected that , for @xmath242 , the determinant @xmath247 will decrease as the eigenvalues of @xmath241 vanish . so , we attempt to minimize the function @xmath248 , which plays the role of a smooth and concave @xcite surrogate of the rank . the optimization problem resulting from the application of the log - det heuristic to ( [ eq : rnkproblem ] ) is a minimization of a concave function over a convex set , and it is thus non - convex . in order to obtain a related convex optimization problem , the log - det objective function is expanded ( to the first order ) in a taylor series about a fixed @xmath249-dimensional matrix @xmath250 , in such a way as to obtain the linear approximation @xmath251\ ] ] where we have used that , for @xmath252 , @xmath253 @xcite . then , dropping the ( irrelevant ) constant terms , we end up with the sdp @xmath254\nonumber\\ \mbox{subject to}&\quad \bm{\mathfrak{c}}\geq 0\,,\quad \tr_2[\bm{\mathfrak{c}}]={\mathbf{1}}_d\quad\mbox{and}\quad \tr_1\left[(\bm{\rho}^{\sf t}\otimes { \mathbf{1}}_d)\bm{\mathfrak{c}}\right]=\overline{\bm{\rho}}\,.\label{eq : logdetheuristicproblemlinear}\end{aligned}\ ] ] the minimum of @xmath248 can be approximated by iteratively solving the sdp ( [ eq : logdetheuristicproblemlinear ] ) , taking for @xmath255 the resulting @xmath241 of the problem solved with input matrix @xmath250 . if we set @xmath256 as the input for the first iteration , the resulting optimization problem coincides with that obtained with the trace heuristic , in which case a rank reduction may not occur in the first iteration . for successive steps , though , rank reduction is generally observed , justifying the initial claim that the log - det heuristic is a refinement of the trace heuristic . as an illustration of the effectiveness of the log - det heuristic , let us design economical cptp maps to transform a given input state of purity @xmath4 into an x - mems wrt @xmath4 , denoted by @xmath118 ( density matrix obtained by plugging ( [ eq : gammafxnp ] ) , ( [ eq : propprimalsol ] ) and ( [ eq : defsfandg ] ) into ( [ eq : rhoxprime ] ) ) . for the input state , we take the following @xmath1-qubit density matrix of purity @xmath257 : @xmath258 where @xmath259 are the ( totally symmetric ) @xmath1-qubit dicke states of @xmath82 excitations @xcite , defined in the computational basis as @xmath260 with the summation running over every distinct permutation of the sequence of @xmath82 ones and @xmath261 zeros . using sedumi , many iterations of the sdp ( [ eq : logdetheuristicproblemlinear ] ) are solved to produce a low rank choi - jamiokowski matrix that maps @xmath262 to @xmath263 . figure [ fig : rankdrop ] shows the evolution of @xmath264 as the iterations progress , where @xmath265 denotes the optimal choi - jamiokowski matrix found at each step . for completeness and comparison , we also plot the obtained optimal values of the sdp ( [ eq : logdetheuristicproblemlinear ] ) at each iteration . our numerical simulations were run with @xmath266 and @xmath267 , which corresponds to the cptp map that collapses the entire bloch ball into the point corresponding to @xmath118 , cf . ( [ eq : sillymm ] ) . plots [ fig : rankdrop]a and [ fig : rankdrop]b correspond to the cases @xmath115 and @xmath268 , respectively . and @xmath267 . plots ( a ) and ( b ) correspond to the cases @xmath115 and @xmath268 , respectively . tick marks are shown on the iterations where a rank drop takes place . for ease of visualization , the optimal values of the rank of @xmath241 and of the objective function of ( [ eq : logdetheuristicproblemlinear ] ) are presented from the third iteration onwards . for plot ( a ) , the rank of the initial choi - jamiokowski matrix is @xmath269 , which decays to @xmath270 during the first and second iterations and reaches the minimum @xmath271 after @xmath272 iterations . for plot ( b ) , the initial rank is @xmath273 , decaying to @xmath274 during the first and second iterations . after @xmath275 iterations the rank reaches @xmath276 and no further decay is observed . ] in the case @xmath115 , plot [ fig : rankdrop]a shows that @xmath277 is reached on the @xmath272th iteration . from there onwards , every produced @xmath278 determines a cptp map that can be written with only two kraus operators . remarkably , in this case , the log - det heuristic provides _ minimal _ kraus decompositions , since no further decrease can occur ( as @xmath279 would correspond to a unitary map ) . in the case @xmath268 , plot [ fig : rankdrop]b shows that @xmath280 is reached after @xmath275 iterations . although not shown in the plot , we ran @xmath281 iterations more and , even so , no further decrease of @xmath282 was observed . of course , that does not mean that a cptp map that implement @xmath283 with two kraus operators does not exist , but merely indicates that if it does exist , then it can not be reached with the log - det heuristic . we conclude this section by noting that , although the log - det heuristic may not always lead to minimal kraus decompositions , it is at least very effective in producing cptp maps that implement a desired state transformation with a relatively small number of kraus operators . it should be noted , though , that the efficiency of the rank decay scheme is significantly limited by the number of qubits involved , as the size of problem ( [ eq : logdetheuristicproblemlinear ] ) scales exponentially with @xmath1 . the characterization of multiqubit mems is a long - standing open problem in quantum information science . in its core , lies the inherent difficulty in quantifying the genuine multipartite entanglement of multiqubit systems @xcite . notwithstanding , as progresses start to be made in this field @xcite , some preliminar sketches of how @xmath1-qubit mems look like can be drawn @xcite . in this paper , we rely on a recently obtained closed formula for the gm - concurrence of @xmath1-qubit @xmath284-states @xcite to determine amongst the set of @xmath1-qubit x - states those with maximal gm - concurrence for ( i ) a fixed set of eigenvalues and ( ii ) for a fixed mixedness ( as measured by purity ) , which we refer to as `` x - mems wrt spectrum '' and `` x - mems wrt purity '' , respectively . using only elementary algebra , explicit forms of density matrices and maximal gm - concurrence were obtained for x - mems wrt every possible spectrum . besides , the unitary transformation that takes an arbitrary @xmath1-qubit state into the corresponding x - mems wrt spectrum was characterized , generalizing a previous result of verstraete _ et al . _ @xcite for two - qubits . although x - mems wrt purity had already been identified in @xcite , we relied on the fact that they form a subset of x - mems wrt to spectrum to numerically reconstruct them and rigorously prove their optimality via sdp methods . additionally , we formulated as a rank minimization problem the design of quantum operations that implements any desired quantum state transformation with the minimal number of kraus operators . then , applying a heuristic method to specific examples of this optimization problem with @xmath115 and @xmath285 qubits , we efficiently characterized low rank quantum operations that transform three- and four - qubit states into the corresponding x - mems wrt purity . an extension of our sdp approach to characterize extreme x - states wrt other measures of mixedness ( e.g. von neumann entropy ) and/or other measures of multipartite quantum correlations / nonclassicality ( e.g. gm - negativity @xcite , global quantum discord @xcite or the measure introduced in @xcite ) is an interesting line for future research . for any desired measures of correlation ( @xmath286 ) and mixedness ( @xmath287 ) , the corresponding extreme x - states are , formally , the optimal points of the problem @xmath288 where @xmath289 specifies a desired value of mixedness and the optimization runs over all x - density matrices @xmath28 . although linearity of @xmath286 and @xmath287 would promptly guarantee that problem ( [ eq : optprblmgen ] ) is a sdp , such a form can also be established in certain non - linear cases . indeed , in this paper , we have seen that despite the specific non - linearities of @xmath286 ( taken as the gm - concurrence ) and @xmath287 ( taken as the purity ) , an equivalent sdp was built by suitably parametrizing @xmath290 ( cf . ( [ eq : eigscgenqubitxstate ] ) ) and applying a standard trick to turn @xmath291 into a lmi ( cf . [ app : quadtolmi ] ) . it is thus conceivable that a similar approach can handle other choices of non - linear measures @xmath286 and @xmath287 . however , determining whether ( and how ) problem ( [ eq : optprblmgen ] ) can be cast as a sdp is expected to strongly depend on the particular choice of measures and , as such , should be considered case by case . along these lines , a particularly interesting problem is to consider a pair of optimization problems formed by fixing @xmath286 as some correlation measure and setting @xmath287 as ( i ) purity and ( ii ) von neumman entropy . the question to be answered here is whether the two problems yield the same set of extreme x - states . such an analysis has been conducted by wei _ _ in the bipartite case with @xmath286 set as concurrence , negativity and relative entropy of entanglement @xcite . remarkably , different extreme states were obtained by changing the mixedness measure , which suggests that the same would occur in the multipartite setting . however , an explicit verification of this conjecture remains an open problem . the authors are indebted to marcus huber and mart perarnau - llobet for bringing reference @xcite to our attention , and to an anonymous referee for valuable comments on a previous version of this manuscript . pemfm thanks g j milburn for discussions and the financial support jointly provided by the brazilian air force and by the program `` cincia sem fronteiras '' , project no . 200024/2014 - 0 . smhr acknowledges financial support from national science foundation grant no . phy-1203931 . in this appendix we briefly review a standard trick to convert quadratic inequality constraints into linear matrix inequalities ( lmi ) . in our proof of theorem [ thm : xmemswrtpurity ] , this was used to write the ( convex ) quadratically constrained linear program ( [ eq : optprblm3 ] ) in the sdp inequality form ( [ eq : optprblm4 ] ) . we illustrate the technique in this particular case . first , we use the well - known formula @xmath292 to rewrite the quadratic constraint in ( [ eq : optprblm3 ] ) as @xmath293 or , equivalently , @xmath294 where @xmath181 and @xmath182 denote the `` all - one '' @xmath29-dimensional vector and matrix , respectively , and @xmath295 we note that @xmath296 is a positive non - singular matrix with eigenvalues @xmath13 ( @xmath297-fold degenerate ) and @xmath298 ( non - degenerate ) , and its inverse is given by @xmath299 now , we recognize the lhs of inequality ( [ eq : quadform ] ) as the schur complement of the following ( block ) matrix of dimension @xmath298 : @xmath300\,.\ ] ] since @xmath301 , we conclude that ( [ eq : quadform ] ) is equivalent to the constraint that matrix ( [ eq : schurcompl ] ) is positive semidefinite @xcite . as a result , problem ( [ eq : optprblm3 ] ) takes the form : @xmath302\geq 0\ ] ] where @xmath303 . up to a constant term and an inversion of sign in the objective function ( see footnote at page ) , this is equivalent to problem ( [ eq : optprblm4 ] ) . 10 url # 1#1urlprefix[2][]#2 greenberger d m , horne m a and zeilinger a 1989 going beyond bell s theorem _ bell s theorem , quantum theory , and conceptions of the universe _ ed kafatos m ( kluwer academics , dordrecht , the netherlands ) pp 6972 . available from http://arxiv.org/abs/0712.0921[arxiv:0712.0921 [ quant - ph ] ] dunkl c f and slater p b 2015 separability probability formulas and their proofs for generalized two - qubit @xmath284-matrices endowed with hilbert - schmidt and induced measures . available from http://arxiv.org/abs/1501.02289[arxiv:1501.02289 [ quant - ph ] ] namkung m , chang j , shin j and kwon y 2014 revisiting quantum discord for two - qubit @xmath284 states : error bound to analytical formula . available from http://arxiv.org/abs/1404.6329[arxiv:1404.6329 [ quant - ph ] ] maldonado - trapp a , hu a and roa l 2015 analytical solutions and criteria for the quantum discord of two - qubit @xmath284-states . available from http://arxiv.org/abs/1501.06900[arxiv:1501.06900 [ quant - ph ] ] mendona p e m f , marchiolli m a and milburn g j 2015 heuristic for estimation of multiqubit genuine multipartite entanglement . available from http://arxiv.org/abs/1501.07026[arxiv:1501.07026 [ quant - ph ] ] huber m , perarnau - llobet m , hovhannisyan k v , skrzypczyk p , klckl c , brunner n and acn a thermodynamic cost of creating correlations . available from http://arxiv.org/abs/1404.2169[arxiv:1404.2169 [ quant - ph ] ] fazel m , hindi h and boyd s 2003 log - det heuristic for matrix rank minimization with applications to hankel and euclidean distance matrices _ proceedings of the american control conference _

for every possible spectrum of @xmath0-dimensional density operators , we construct an @xmath1-qubit x - state of same spectrum and maximal genuine multipartite ( gm- ) concurrence , hence characterizing a global unitary transformation that constrained to output x - states maximizes the gm - concurrence of an arbitrary input mixed state of @xmath1-qubits . we also apply semidefinite programming methods to obtain @xmath1-qubit x - states with maximal gm - concurrence for a given purity and to provide an alternative proof of optimality of a recently proposed set of density matrices for the purpose , the so - called x - mems . furthermore , we introduce a numerical strategy to tailor a quantum operation that converts between any two given density matrices using a relatively small number of kraus operators . we apply our strategy to design short operator - sum representations for the transformation between any given @xmath1-qubit mixed state and a corresponding x - mems of same purity . _ keywords _ : entanglement , genuine multipartite concurrence , @xmath1-qubit x - states

the study of systems of @xmath0-level atoms is interesting as proposals have been made to use them as quantum memories or to manipulate quantum information , among other applications . cavity qed systems in particular have been favored because of their advantage when subjected to coherent manipulations , and schemes have been presented for various quantum gates using 3-level atoms and trapped ions @xcite . in the presence of an electromagnetic field , few - level atomic systems present the phenomenon of _ superradiance _ @xcite where the decay rate is proportional to the square of the number of atoms , @xmath2 , instead of @xmath3 ( the expected result for independent atom emission ) . there thus exists a _ separatrix _ in parameter - space which divides a so - called _ normal _ region from a region of collective behavior @xcite . in this contribution we study a @xmath0-level atomic system in the @xmath1-configuration ( cf . [ f1 ] ) interacting through a one - mode electromagnetic field , and we show that its phase diagram for the ground state presents a triple point independently of the number of atoms , which prevails in the thermodynamic limit . we use the fidelity and the fidelity susceptibility of neighboring states to determine the sudden changes , in parameter space , in the ground state composition . we also find that , when in double resonance , there is a mirror symmetry in the energy spectrum around the energy value @xmath4 , where @xmath5 is the total excitation number ( which turns out to be a constant of motion in the rotating - wave approximation ( rwa ) ) . , or _ ladder _ , the @xmath6-th level frequency is denoted by @xmath7 with the convention @xmath8 , and the coupling parameter between levels @xmath6 and @xmath9 is @xmath10 . ] the intrinsic hamiltonian for a @xmath0-level system in the rwa is @xcite @xmath11 \nonumber \end{aligned}\ ] ] where @xmath12 are the creation and annihilation operators of the field , @xmath13 are the atomic operators , the @xmath6-th level frequency is denoted by @xmath7 with the convention @xmath8 , and the coupling parameter between levels @xmath6 and @xmath9 is @xmath10 . the detuning between levels @xmath6 and @xmath9 and the field is denoted by @xmath14 , where @xmath15 is the field frequency . it has @xmath16 constants of motion , viz . , the total number of atoms @xmath17 , and the total number of excitations @xmath18 . as the system is non - integrable , one may solve via numerical diagonalization . a natural basis in which we diagonalize our hamiltonian is @xmath19 @xcite . here , @xmath20 represents the number of photons of the fock state , @xmath21 and @xmath22 are the atomic population of levels @xmath23 , respectively . notice that the hamiltonian ( [ hamiltonian ] ) is invariant under the transformation @xmath24 and @xmath25 , which preserves the commutation relations of the bosonic operators . for this reason we consider only positive values for @xmath10 . the energy of the ground state as a function of the coupling parameters @xmath26 and @xmath27 is displayed in fig . [ f2]a , in double resonance . we use @xmath28 for clarity of the figure . the intersection lines between @xmath29 with @xmath30 from @xmath31 are also shown . all these _ separatrices _ may be determined by using the fidelity @xmath32 or the fidelity susceptibility @xmath33 of neighboring states @xcite ( for a complete review of the use of these concepts see @xcite ) , defined by @xmath34 & & \chi = 2\,\frac{1-f(\lambda,\,\lambda+\delta\lambda)}{(\delta\lambda)^2}.\end{aligned}\ ] ] a ) . the intersection lines between @xmath29 with @xmath30 from @xmath35 are also shown . at @xmath36 the regions @xmath37 and @xmath38 meet : a true triple point . b ) a cut at @xmath39 is shown for @xmath40 ( lower , blue line ) and @xmath41 ( upper , red line).,title="fig : " ] + b ) . the intersection lines between @xmath29 with @xmath30 from @xmath35 are also shown . at @xmath36 the regions @xmath37 and @xmath38 meet : a true triple point . b ) a cut at @xmath39 is shown for @xmath40 ( lower , blue line ) and @xmath41 ( upper , red line).,title="fig : " ] the fidelity is a measure of the distance between states , for a pure state @xmath42 which varies as a function of a control parameter @xmath43 . the fidelity susceptibility , essentially its second derivative with respect to the control parameter , is a more sensitive quantity . figures [ f3]a , b show the zeroes attained by the fidelity measure and the divergences in the fidelity susceptibility at each separatrix , where the value of the excitation number @xmath5 changes . a cut along the straight line @xmath44 in parameter space was taken for good comparison with fig . [ f2]a . here , @xmath28 as well . ( note that a line of the form @xmath45 is not a convenient choice , as the parameter would factorize in the interaction hamiltonian and the eigenstates would remain the same as @xmath26 changes within a region . ) whereas the fidelity measure clearly marks the transition points , it fails at distinguishing states within a region of constant @xmath5 ; the fidelity susceptibility , being a much more sensitive function , does distinguish them . a ) between neighboring states , along the line @xmath44 in parameter space , plotted as a function of @xmath27 . it is zero at the transition points . b ) fidelity susceptibility @xmath33 as a function of the coupling parameter @xmath27 ; it diverges at each point where the excitation number @xmath5 changes.,title="fig : " ] + b ) between neighboring states , along the line @xmath44 in parameter space , plotted as a function of @xmath27 . it is zero at the transition points . b ) fidelity susceptibility @xmath33 as a function of the coupling parameter @xmath27 ; it diverges at each point where the excitation number @xmath5 changes.,title="fig : " ] the innermost _ separatrix _ divides the space into a _ normal _ region , where the number of excitations @xmath5 vanishes , and a _ collective _ region . this separatrix prevails in the thermodynamic limit @xmath46 . at @xmath36 the regions @xmath37 and @xmath38 meet : a _ triple point _ which is fixed for all values of the number of atoms @xmath3 . all other triple points as well as all other separatrices slide towards the boundary between @xmath47 and @xmath48 as @xmath3 increases , and new transition regions intersecting the region @xmath47 appear . this holds true all the way to the thermodynamic limit , where all separatrices coalesce into one . a manifestation of this phenomenon is shown in fig . [ f2]b , where a cut at @xmath39 of the ground state energy is shown for @xmath40 and @xmath41 ; the dots along the lines show where the transitions from one @xmath5 to the next take place along the @xmath26-axis . as @xmath3 increases , the intersection points tend to @xmath49 . the same holds true for any other value of @xmath27 . the triple point itself is shown in fig . [ f4 ] , together with a diagramatic description of the ground states corresponding to each value of @xmath5 , for @xmath28 . this means that in a vicinity of this point any quantum fluctuation will drastically change the composition of the ground state . since the dimension of the hilbert space does not change for larger values of @xmath3 , the same composition as that shown in the figure is obtained for a larger number of atoms ( in fact , since the number of excitations is low , increasing @xmath3 amounts to increase the occupation of the first level ) . it is worth stressing that the existence of this triple point _ independent _ of @xmath3 is a characteristic of the @xmath1-configuration ; it does not appear in the @xmath50 or the @xmath51 configurations . . the figure shows the transformation of the ground state near the vicinity of this point due to quantum fluctuations , for @xmath28 . the triple point persists for _ any _ @xmath3 , including the thermodynamic limit . ( wavy lines represent photons and an `` @xmath52 '' represents a level occupation . ) ] the energy spectrum at the triple point for a system of @xmath53 particles is displayed for all values of @xmath5 up to @xmath54 in figure [ f5]a , for the double resonant case . when the energy is plotted against the excitation number @xmath5 we observe a mirror symmetry with respect to @xmath4 , and the symmetric states correspond to eigenstates whose components differ only by phases . the spectrum in fig . [ f5]b shows this for @xmath53 . this means that all expectation values of operators will be equal for these symmetric states , except for the energy itself . a ) particles at the triple point . the @xmath5 values run from @xmath55 to @xmath54 . b ) energy spectrum as function of the total excitation number for @xmath53 at the triple point . there is a mirror symmetry with respect to @xmath4 ( indicated by dots ) for each value of @xmath5.,title="fig : " ] + b ) particles at the triple point . the @xmath5 values run from @xmath55 to @xmath54 . b ) energy spectrum as function of the total excitation number for @xmath53 at the triple point . there is a mirror symmetry with respect to @xmath4 ( indicated by dots ) for each value of @xmath5.,title="fig : " ] the system presents degenerate states only at @xmath4 , which is not always attained . for instance , if @xmath5 is odd and @xmath56 there are no states with @xmath4 . for @xmath57 there is a degeneracy of @xmath58 , where @xmath59 denotes @xmath60 . for other values it depends strongly on @xmath5 . for small values of the total excitation number @xmath5 the dimension of the hilbert space depends solely on @xmath5 , and not on @xmath3 . this allows one to study the system for a large number of atoms , including the thermodynamic limit . in this limit @xmath61 , the energy spectrum becomes independent of @xmath27 and shows a collapse of energy levels for all values of @xmath5 at the point @xmath62 . here , the lowest energy levels for all @xmath5 have a value @xmath63 , the next lowest energy levels for all @xmath5 have a value @xmath64 , the next levels have @xmath65 , and so on , as shown in figure [ f6]a where the energy spectrum is plotted against the coupling parameter @xmath26 for values of @xmath5 from @xmath55 to @xmath66 . the spectrum as a function of @xmath5 becomes equidistant with only even harmonics in this limit ( cf . a chain of excited - state quantum phase transitions demarcating the superradiant phase was previously reported for the jaynes - cummings and dicke models of quantum optics @xcite . it was shown that the emergence of quantum chaos is caused by the precursors of the excited - state quantum phase transitions . finite - size scaling behavior of excited - state quantum phase transitions at the mean - field level was studied for 2-level bosonic and fermionic models in @xcite . a ) ) as a function of @xmath26 from @xmath47 to @xmath67 . one can see the collapse of the successive energy levels for each value of @xmath5 at the energy values @xmath68 at the triple point . b ) energy spectrum as a function of @xmath5 at the triple point . notice that for all values of @xmath5 we have an equidistant spectrum with only even harmonics.,title="fig : " ] + b ) ) as a function of @xmath26 from @xmath47 to @xmath67 . one can see the collapse of the successive energy levels for each value of @xmath5 at the energy values @xmath68 at the triple point . b ) energy spectrum as a function of @xmath5 at the triple point . notice that for all values of @xmath5 we have an equidistant spectrum with only even harmonics.,title="fig : " ] one can also calculate the expectation values of the population operators @xmath69 and @xmath70 at the triple point , for each of the eigenstates of the hamiltonian ( [ hamiltonian ] ) figure [ f7 ] shows these for @xmath71 and @xmath67 . we again see the inherited mirror symmetry with respect to the middle state . ( continuous , red ) , @xmath72 ( dashed , blue ) , and the number of photons @xmath20 ( dotted , green ) , for each eigenstate of the hamiltonian . here , @xmath73 . ] the density matrices of the eigenstates @xmath74 and @xmath75 , where @xmath74 labels the states in order of increasing energy and @xmath76 is the dimension of the hilbert space of the system , have the same coefficients and their coherences differ only by a phase , except for the degenerate states @xmath4 . a ) of having @xmath77 photons in the system at the triple point in double resonance , as the excitation number @xmath5 grows without limit.,title="fig : " ] + b ) of having @xmath77 photons in the system at the triple point in double resonance , as the excitation number @xmath5 grows without limit.,title="fig : " ] another interesting limit is @xmath78 , where the dimension of the hilbert space depends only on @xmath3 @xcite . for a fixed number @xmath3 of atoms , as @xmath5 increases the atomic excitations may saturate and the population of photons grows . it is therefore interesting to study the photon distribution in the system at the triple point . the probability of having a certain number @xmath20 of photons is obtained by taking the modulus squared of the scalar product of the fock state @xmath79 and the eigenstate . given @xmath3 and @xmath5 , the minimum number of photons in the qed cavity is @xmath80 . we consider here the ground state at each value of the excitation number @xmath5 and evaluate the excess number of photons , i.e. , the probability @xmath81 of having @xmath82 photons in the system . this is plotted in figure [ f8]a for @xmath83 atoms and for @xmath84 and the limit @xmath85 . the arrows in the figure show how each value moves as @xmath5 grows without limit . both distributions are quasi - gaussian , though this is not very clear in the figure . for a larger number of atoms , @xmath86 , fig . [ f8]b shows better the photon probability distributions , this time for @xmath87 and @xmath85 . a triple point in the phase diagram of a three - level system interacting with a one - mode electromagnetic field is found for the atomic @xmath1-configuration . this triple point is fixed in parameter space , is present for any finite number of atoms , and prevails in the thermodynamic limit . when in double resonance it resides at @xmath36 . away from the double resonance condition it acquires different coordinates , but it is still fixed as @xmath88 . properties of observables at the triple point were studied , including state structure , atomic level populations , energy spectra , and photon number probabilities . we find that the energy spectrum has a mirror symmetry with respect to the point @xmath4 , where @xmath5 is the total excitation number ( a constant of motion of the system ) .

the energy spectrum of a @xmath0-level atomic system in the @xmath1-configuration is studied . this configuration presents a _ triple point _ independently of the number of atoms , which remains in the thermodynamic limit . this means that in a vicinity of this point any quantum fluctuation will drastically change the composition of the ground state of the system . we study the expectation values of the atomic population of each level , the number of photons , and the probability distribution of photons at the triple point .

anomalous x - ray pulsars ( axps ) and soft gamma repeaters ( sgrs ) are part of a new class of neutron stars ( nss ) , whose emission mechanisms commute with neither accreting x - ray pulsars nor radio pulsars ( psrs ) . their periods are clustered between 2 - 12 s and they are characterized by their rapid spin down ( @xmath2 ) and high quiescent luminosities ( @xmath3 ) that can not be explained by spin down mechanisms . they also show wide range of variability ; including flux and pulse profile variations ( dib , kaspi & gavriil 2007a ; tam et al . 2008 ) , glitches ( dib , kaspi & gavriil 2007b ) , bursts ( gavriil , kaspi & woods 2002 ) and strong timing noise ( baykal et al . 2000 ; kaspi , gavriil & chakrabarty et al . three sgrs are known to exhibit giant bursts ( sgr 0526 - 66 , sgr 1900 + 14 , sgr 1806 - 20 ) , which in two of these objects rapid quasi - periodic oscillations ( qpos ) are detected after the onset of their flares ( sgr 1900 + 14 ; strohmayer & watts 2005 , sgr 1806 - 20 ; israel et al . 2005 ) . sgrs and axps are widely accepted as being magnetars , neutron stars ( nss ) with magnetic fields strengths greater than the quantum critical level ( @xmath4 gauss ) , whose main source of energy is the dissipation of these fields ( duncan & thompson 1992 ; thompson & duncan 1995 ) . the magnetar model for these objects explains the sgr bursts and the large flares , and it predicted the axp bursts which occurred in 1e 1048.1 - 5937 ( gavriil , kaspi & woods 2002 ) and 1e 2259 + 586 ( gavriil & kaspi 2002 ) . thompson & duncan ( 1995 ) argue that the bursts are due to crustal cracking as a result of magnetic diffusion , and that the large flares are due to major reconfigurations of the global magnetic field ( thompson & duncan 1995 ) . magnetar strength magnetic fields can also account for the rapid spin down of these objects and the qpos observed during large flares ( e.g , israel et al . 2005 ; strohmayer & watts 2005 ) . the existence of a fossil disk , and accretion from it , has also been proposed to explain the persistent properties of these sources ( van paradijs et al . 1995 ; chatterjee et al . 2000 ; alpar 2001 ) . however , this model has difficulty explaining the giant flares , repeat bursts and the pulsed optical emission that exists in this class of objects . in this paper , we investigated the properties of the broad - band noise in selected axps and sgrs , and present plausible origins for it within the context of the above two models . for example , if axps and sgrs have fall - back disks that they accrete from , as has been suggested by ertan et al . ( 2007 ) , then we can compare this noise to that of other accreting sources . for instance , the broad - band noise in x - ray binaries spans a very broad range and is due to variable x - ray flux from the accretion disk . in corona hosting x - ray binaries , a contribution of the noise in the high frequency range is due to photon delays in compton up scattered radiation ( van der klis 2006 and references therein ) . this is important given that axps and sgrs show hard power law like tails in their energy spectra that are also most likely due to photons being up - scattered in energy . in the context of the magnetar model , these hard - tails in axps and sgrs have been successfully modeled by resonant cyclotron scattering ( lyutikov & gavriil 2006 ; rea et al . we argue that resonant cyclotron scattering can also introduce broad band noise in the frequency range observed for these sources . thus , in order to discriminate between the two models it becomes of great interest to study the broad - band noise in axps and sgrs . llccl object & obs i d & ave . count rate & burst@xmath5 + & & c / s & + 1e 1048.1 - 5937 & 40083 - 08 & 3.889 & no + 1e 1048.1 - 5937 & 50082 - 04 & 3.771 & no + 1e 1048.1 - 5937 & 60069 - 03 & 3.302 & yes + 1e 1048.1 - 5937 & 70094 - 02 & 7.030 & no + 1e 1048.1 - 5937 & 80098 - 02 & 5.667 & no + 1e 1048.1 - 5937 & 90076 - 02 & 3.580 & yes + 1e 1048.1 - 5937 & 91070 - 02 & 3.255 & no + 1e 1048.1 - 5937 & 92006 - 02 & 3.457 & no + 1e 2259 + 586 & 20145 - 01 & 1.799 & no + 1e 2259 + 586 & 20146 - 01 & 2.269 & no + 1e 2259 + 586 & 40082 - 01 & 1.752 & no + 1e 2259 + 586 & 50082 - 01 & 1.167 & no + 1e 2259 + 586 & 60069 - 01 & 0.4600 & no + 1e 2259 + 586 & 70094 - 01 & 2.695 & yes + 1e 2259 + 586 & 80098 - 01 & 1.765 & no + 1e 2259 + 586 & 90076 - 01 & 1.673 & no + rxs j1708 - 40 & 40083 - 14 & 13.02 & no + rxs j1708 - 40 & 50082 - 07 & 12.10 & no + rxs j1708 - 40 & 60069 - 07 & 12.37 & no + rxs j1708 - 40 & 60412 - 01 & 12.99 & no + rxs j1708 - 40 & 70094 - 04 & 13.58 & no + rxs j1708 - 40 & 80098 - 04 & 13.38 & no + rxs j1708 - 40 & 90076 - 04 & 13.38 & no + 1e 1841 - 045 & 40083 - 11 & 9.521 & no + 1e 1841 - 045 & 50082 - 05 & 8.835 & no + 1e 1841 - 045 & 70094 - 03 & 9.937 & no + 1e 1841 - 045 & 90076 - 03 & 9.013 & no + 1e 1841 - 045 & 91070 - 03 & 9.006 & no + sgr 1806 - 20 & 50142 - 01 & 7.626 & yes + sgr 1806 - 20 & 50142 - 03 & 7.607 & yes + sgr 1806 - 20 & 60121 - 01 & 7.053 & yes + sgr 1806 - 20 & 70136 - 02 & 10.69 & yes + sgr 1806 - 20 & 90073 - 02 & 10.93 & yes + sgr 1806 - 20 & 90074 - 02 & 12.70 & yes + sgr 1806 - 20 & 91062 - 02 & 8.721 & yes + + + + to calculate the band limited noise of these objects we used the 2 - 60 kev archived public rossi x - ray timing explorer ( rxte ) light curves . the rxte archive has a total of 10 sources that are either axps or sgrs . we used only appropriate sets of data , eliminating sources depending on epochs , exposures , burst dates and objects in the field of view ( fov ) . we recovered 1 known sgr and 4 known axp observations with fields relatively free of other sources . the exceptions are 1e 1048.1 - 5937 with eta carinea and sgr 1806 - 20 with xte j1810 - 197 in their fov , which both do not effect the noise levels in our analysis . the variability of eta carinae occurs in a very different time scale ( @xmath185 days ; corcoran et al . 1997 ) with respect to 256 seconds used in this study . xte j1810 - 197 is a transient source which was quiescent during the observations in our analysis ( ibrahim et al . other fields include strong and variable sources . the data were obtained by the proportional counter array ( pca ; jahoda et al . 1996 ) instrument aboard rxte . we obtained 125 ms resolution background subtracted light curves from archived standard products ( stdprods ) of rxte . these light curves are screened through standard procedures ; namely source elevation greater than @xmath6 , pointing offset smaller than @xmath7 , also pcu turn - off times are subtracted through defining good time intervals . backgrounds are generated by pcabackest command of ftools with the use of background models of the relevant epoch . these net lightcurves also include the barycentrically corrected time columns , which were used in our analyses . only when necessary , particularly to check background effects , non - background subtracted light curves were created from original data using seextrct v4.2 . the manipulation of the data was made with the ftools v5.21 software . in order to measure the broad - band noise , we derived and averaged several power spectra for each source . the power spectral densities ( psd ) expressed were calculated in terms of the fractional rms amplitude squared following miyamoto et al . ( 1991 ) , and the expected white noise levels were subtracted hence leaving us with the rms fractional variability of the time series in units of ( @xmath8/hz . we used in total 658 light curves for our analysis ( 151 for 1e 2259 + 586 , 136 for 1e 1048.1 - 5937 , 125 for sgr 1806 - 20 , 82 for 1e 1841 - 045 and 153 for rxs j1708 - 40 ) , each of which was segmented into 1050 individual power spectra to construct average psd in a given time line . in order to increase the signal to noise of the psd to look for broad - band noise , the averaged light curves are grouped in three month intervals yielding 20 ks50 ks long exposure times for each source . _ the burst intervals noted in table 1 are excluded from our data samples to avoid additional red noise because of the light curve variability during bursts . the remaining burst like features ( especially for the bursting sgr 1806 - 20 ) and spurious events due to instrument errors are screened by hand from the light curves . _ we averaged several 256 s long psd for each source . figures 1 through 4 show the representative averaged psd fitted with a broken power law model ; @xmath9^{-\alpha}$ ] , in which @xmath10 is the frequency , @xmath11 is the ( @xmath8/hz value of the flat part , @xmath12 is the cutoff frequency and @xmath13 is the power law index . each spectra in the analyses is averaged over three - month intervals and are separated by 5900 days . in all sources , we find red noise with a break in the averaged psd which may be considered band limited noise ( bln ) . for all figures , the reduced chi squares of the fits are between 0.6 - 1.2 , the range of observed break frequencies are between 0.05 - 0.7 hz and the power law indices are between ( -)0.5 - 375 . table 2 shows the list of power law indices derived from the fits to psd of the sources in time . an irregular variation of the indices in time is apparent . the ranges in table 2 denote 1 @xmath14 confidence level error ranges from fits . it is very hard to determine the value of these indices , since the decline in the power spectrum is near the edge of the sampling window and errors also increase considerably . we would like to make a note that for 2 @xmath14 errors on the power law index , the canonical value of -2 is in the acceptable range . in addition , the fitted psd are integrated over @xmath15 hz interval to get the integrated fractional rms values . time evolution of these values for 1e 1048.1 - 5937 and 1e 2259 + 586 are plotted in figure 5 . we found that for 1e 1048.1 - 5937 and 1e 2259 + 586 ( two axps ) , their burst ( maybe also flare ) epoch correlate with their broad - band noise levels . the steady @xmath16 rms noise of 1e 1048.1 - 5937 showed a slight drop , during which two of its sgr - like short bursts and a flare appeared , to levels around @xmath17 rms in 16 months . this low noise recovered throughout the next 5 months until it reached its former noise level ( see figures 1 and 5 ) . the red noise level of the source rose after a long burst and later a glitch and large scale flux increase ( dib et al . 2007 ; tam et al . 2008 ) to levels @xmath18 rms and stayed at this level ( with some variability ) for about 1.95 years ( see fig . 5 , top panel ) . 1e 2259 + 586 showed a slightly different behavior . it had no detectable broad - band red noise in quiescence and developed broad - band red noise around @xmath19 rms level right after its burst active state and the on - set of a glitch , stayed consistently at this level throughout nearly 2 years , until it diminished in about 1.5 months ( at mjd 53233 ) ( see figures 2 and 5 bottom panel ) back to its original level of no broad - band noise . the other two axps 1e 1841 - 045 and rxs j1708 - 40 maintain a persistent broad - band noise level compared with the two axps discussed in the above paragraph . sources showed around @xmath20 rms and @xmath21 rms level noise for 1e 1841 - 045 and rxs j1708 - 40 , respectively ( with slight increase in time ) . the total time span of the detected broad - band noise for the two different epochs for rxs j1708 - 40 were 2.3 years ( 51215 - 52053 mjd ) and 2.6 years ( 52366 - 53325 mjd ) . also for 1e 1841 - 045 the time spans for two separate time epochs analized in this work were 2 years ( 51977 - 51260 mjd ) and 3.5 years ( 53635 - 52349 mjd ) . 1e 1841 - 045 also stands out as the noisier of the two . sgr 1806 - 20 exhibited broad - band noise slightly fluctuating about a persistent level of @xmath17 , overall . the time span of our analysis in two epoch ranges were 1.4 years ( 51600 - 52100 mjd ) and 1.7 years ( 53000 - 53600 mjd ) . we had difficulty relating the burst activity and the psd for sgr 1806 - 20 due to the existence of excessive number of bursts . .the power law indices of axps and sgrs derived from fitted psd in this work . [ cols="<,<,<,<,<",options="header " , ] figure 5 , shows the comparative evolution of the band - limited red noise in 1e 1048.1 - 5937 and 1e 2259 + 586 . the ordinate is the integrated rms band - limited noise and the abscissa is the time in mjd . the error in time ( x error ) is the total time of the accumulated observation ids for each data point . the y error is the 1 @xmath14 error in the integrated rms values . 1e 2259 + 586 showed the highest level of noise and longest duration of change . 1e 1048.1 - 5937 is the only source that showed a drop of the noise level assuming level at start of mjds is the steady normal level of noise . it was also the most noisy source after 1e 2259 + 586 . a significant rise of the rms broad - band noise level in 1e 1048.1 - 5937 ( like 1e 2259 + 586 ) after the long - burst epoch at 53185 mjd is detected for about 1.95 years ( may have lasted about 3 years judging from telegrams ; rea et al . 2007 ) . on the basis of torque noise , 1e 2259 + 586 is very quiet ( baykal et al . 2000 ; gavriil & kaspi 2002 ) , but after its outburst and a glitch , 1e 2259 + 586 experienced rapid spin down over 60 days ( woods et al . 2004 ) and showed an excess in infrared radiation ( hulleman et al . 2001 ) . in this period 1e 2259 + 586 displayed band - limited noise lasting for a timescale of 2 years and no band - limited red noise other wise . 1e 1048.1 - 5937 shows a highly variable torque noise such that , it can not be phase connected for long time intervals which is consistent with accreting sources ( baykal et al . 2000 ; kaspi , gavriil & chakrabarty et al . 2001 ) . in our work , this source showed band - limited red noise at all times . the substantial increase of the broad - band noise of 1e 1048.1 - 5937 in figure 5 ( from 15@xmath0 to about 25 - 30@xmath0 ) occured gradually after the the long burst ( @xmath22 699 sec ) at 53185 mjd ( gavriil , kaspi & woods 2006 ) . following this , there has been a flux increase ( outburst - like ) from 1e 1048.1 - 5937 reported along with a spin - up glitch at 54186 mjd ( dib et al . 2007 ; tam et al . 2008 ) . the decrease of this rise of the soft x - ray flux was reported to be at 54265 mjd ( rea et al . 2007 ; tam et al . 2008 ) . these latest dates are at the edge of our curve for 1e 1048.1 - 5937 in figure 5 and are not included in our results . however , it is clear that it follows our peak of integrated rms values for the source . _ we note that the rise in broad - band noise started much earlier than the flux outburst right after the long burst at 53185 mjd ( gavriil , kaspi & woods 2006 ) . _ on the other hand , we caution that given the sparsity of this event , and the possibility that similar bursts went unnoticed , this start for the rise in noise is not as clear as in 1e 2259 + 586 . we also note that this increase in red noise is not very correlated with a large increase in persistent and pulsed flux in the early stages until the outburst - like flux increase . prior to its burst at 53185 mjd and the broad - band noise increase , there was a slight decrease of red noise during which two sgr - like short bursts ( 52211 mjd and 52227 mjd ) and two flares ( 52218 mjd and 52444 mjd ) appeared ( gavriil , kaspi & woods 2002 ; gavriil & kaspi 2004 ) . the two bursts showed characteristics similar to sgr bursts like marginal pulsed flux increase not more than 3@xmath14 and short duration ( @xmath2350 s 1st burst and @xmath23 2 s 2nd burst ) . since we did not detect any correlation of broad - band noise with the sgr bursts in sgr 1806 - 20 , we do not expect any effects here . the first flare starts between the two sgr - like bursts at 52218 mjd , after which an epoch of lower rms noise started . the drop of the broad - band noise showed a brief increase in the second flare ( maximum at 52444.4 mjd , see fig . 5 ) and the trend of decrease and recovery of noise after the flare is similar to the torque variability that was observed by tam et al . ( 2008 ) . the time - averaged psd of the four axps and the sgr showed in general very low frequency noise . we detected that this noise to be band - limited and in the frequencies higher than 0.05 hz . there were little or no noise at all times below these frequencies given the sensitivity of rxte ( also background ) and statistical quality of rxte light curves . the broad - band noise of axps and sgrs can be compared with the other isolated pulsars or x - ray binaries on the basis of their origin ( e.g. , accretion process , coronal activity , rotational variation ) . broad - band noise in x - ray binaries is a result of aperiodic x - ray flux variability in the form of continuum noise and qpos with frequencies ranging from several mhz to more than thousand hz ( hasinger @xmath24 van der klis , 1989 , van der klis 2000,2006 ) . the flattening in the lower frequencies resemble psd in disk - fed x - ray binaries , specifically the low hard state ( ls ) noise of black hole candidates and the very low frequency noise ( vlfn ) of low - magnetic field neutron star lmxbs ( van der klis , 2006 and references therein ) . a flat topped noise arises in black hole candidates with large integrated rms noise level ( @xmath25 ) in the low frequencies ( van der klis 2006 ) . the vlfn of lmxbs show the flat topped noise with a few percent integrated rms and the continuum power law component shows indices between ( -)1.5 - 2 , in general a range which holds for all the x - ray binaries ( van der klis 2006 ) . the vlfn of the x - ray binaries is correlated with @xmath26 variations associated with an accretion disk and plausibly related to unsteady nuclear burning ( van der klis 2006 ) . the structure of the flattening in the psd sample of axps and sgrs resembles the vlfn of accreting x - ray binaries . if one assumes that the variable x - ray flux originates from an accretion disk with a size between 1@xmath2710@xmath28 cm to 1@xmath2710@xmath29 cm , ( values obtained from ertan et al . 2007 ) the range of keplerian frequencies ( r@xmath30 = gm@xmath31/4@xmath32 ) that should appear in the continuum noise is then from 0.058 hz to 2@xmath2710@xmath33 hz . this range overlaps with the detected range of broad - band noise in our work which supports that we could be detecting noise from the fall - back accretion disks suggested to exist in these systems ( ertan et al . however , the variations of this noise does not resemble the case for x - ray binary sources ( see figure 5 ) . in x - ray binaries outburst effects last for a few days to a week ( where one detects a change in the broad - band noise level , as well ) . the persistent level of emission changes with accretion rate from high states to low states of x - ray binary sources while one also detects reallocation of broad - band noise from high frequencies into low frequencies . on the other hand , in our entire data set we did not detect any reallocation of noise , the noise we detected always appeared in the 0.05 - 0.005 hz range and our timescale for increased noise or high level of broad - band noise lasted for over a year or two which is not a detected phenomenon in x - ray binaries . as discussed in the above paragraphs we favor that the changes in the band limited low frequency noise in magnetars are related most likely to the burst characteristics . glitches seem to be the least effective of all in the broad - band noise changes from these sources . as discussed in the above paragraph , one origin for the detected broad - band noise in axp and sgrs could be the suggested fall - back disks around these objects with the aforementioned caveats . the origin of this noise could also be partially related to the frequency fluctuations of the neutron stars as in isolated pulsars and also to the variability that arises from the up - scattered radiation in their magnetospheres , particularly enhanced in the presence of a magnetar corona ( beloborodov & thompson 2005 , beloborodov & thompson 2007 ) . the plasma coronae in magnetars are a result of the starquakes and twisting of the magnetosphere that injects an electric current ( i.e. , plasma ) into the stellar magnetosphere ( beloborodov & thompson 2005 , beloborodov & thompson 2007 ) . emerging currents are maintained during the x - ray outburst and a dense thermalized plasma is present in the magnetosphere ( lasting from 1 - 10 years ) . broad - band noise due to processes such as photon delays between soft and hard x - ray photons are well studied in x - ray binary sources , however such timescales are much shorter like a few milliseconds as a result of compton - up scattering ( in the presence of thompson optical depth ) from disk coronae ( e.g. , vaughan et al . 1994 , schulz & wijers 1993 ) which yields a continuum noise in very high frequencies in comparison with the frequency range we detect here ( within 256 sec ) . however , the processes in the magnetosphere of axps and sgrs are more complicated . lyutikov & gavriil ( 2006 ) shows that magnetospheric plasma can distort the thermal x - ray emission emerging from the star surface by repeated scatterings at the cyclotron resonance ( e.g. , resonant cyclotron scattering ) . recently , rea et al . ( 2008 ) showed that an x - ray spectral model that involves resonant cyclotron scattering ( together with a blackbody model ) successfully fits the soft x - ray emission ( 1 - 10 kev data ) of ten magnetars . if the dominant scattering process is the resonant scattering rather than thompson scattering ( as shown by rea et al . 2008 ) , then since the thompson optical depth to scattering is different than resonant scattering optical depth , the escape times of photons from the magnetosphere or an existing magnetar corona will be different in comparison with a simple disk corona where thompson scattering optical depth dominates ( as photons are compton up - scattered in the corona ) . once escape times differ , so will the time delays between soft and hard photons will be different in a resonant scattering dominated plasma . rea et al . ( 2008 ) calculates the ratio of optical depths as @xmath34/@xmath35 = 10@xmath36 which strongly suggests that given similar geometries the high level of resonant optical depth to scattering should cause more delays in the emerging photons . if one assumes that scattering optical depth is directly proportional to number of scatterings , a millisecond delay then should become about 100 sec delay , that could yield noise around 0.01 hz . we suggest that especially for 1e 2259 + 586 this scattering can produce or contribute to the band - limited red noise we detect in the 0.005 - 0.05 hz band and its changes . this is , also , supported by the commensurability of 2 years rise in noise we detect in two axps and the magnetar coronal time scale of 1 - 10 years ( time scale the magnetar corona exists ) . moreover , the increase in the photon indices of power - law tails in the x - ray spectra : for 1e 2259 + 586 , @xmath37 increases from -4.3 to -2.8 after the outburst ( woods et al . 2004 , kuiper et al . 2006 ) ; for 1e 1048.1 - 5937 , @xmath37 increases from -3.4 to -3.0 slowly after the reported burst at 53185 mjd and this increases to -2.3 after the flux outburst in 2007 march ( tiengo et al . 2005 ; campana & israel 2007 , rea et al . 2007 , tam et al . 2008 ) . characteristic of the activity is important for inducing such a process , e.g. a long burst or an outburst like characteristics are necessary to produce such effects of highly increased broad - band noise levels . glitches or flares show no effect or slight variations in noise levels . this supported by nearly stable band limited noise levels of 1rxs j1708 - 40 and 1e 1841 - 045 which show only glitches and no bursts . pulsar / magnetar / main.html ] moreover , isolated pulsars that show glitches do not show such long term significant rises in red noise levels in the low frequency band . other models explaining the magnetar timing variabilities are ruled out mostly because of their time scales . for example , seismic vibrations that are used to explain qpos are very rapid and connected to giant bursts , and thus is not a satisfactory explanation for broad - band noise . one modifying principle for the seismic models is the propagation of alfvn waves towards the core . levin ( 2006 ) noted that these oscillations are a continuum , but they decay rapidly thus effectively they can not correspond to our persistent rms noise observations . we observed broad - band noise ( bln ) in the four axps ( 1e 1048.1 - 5937 , 1rxs 1708 - 40 , 1e 1841 - 045 and 1e 2259 + 586 ) and an sgr ( sgr 1806 - 20 ) mainly over the 0.005 - 0.05 hz frequency band . the integrated rms values were in a range 2.5 - 15@xmath0 for four sources and 0@xmath0 rms for 1e 2259 + 586 before bursts . after bursts ( our data do not include burst / outburst epochs ) 20 - 70@xmath0 integrated rms noise was observed in two cases ; 1e 2259 + 586 and 1e 1048.1 - 5937 , respectively . the broad - band noise characteristics do not resemble either the accreting x - ray binaries or the isolated pulsars completely . in general , we found that the four sources indicate a persistent level of integrated rms in time over a long base line of 3.55.5 years which may be affected by noise from frequency fluctuations , and most likely as a result of variability due to resonant cyclotron scattering in a magnetar coronae and/or variability from existing fall - back disks around these objects . we also detected large scale changes of the broad - band noise particularly due to long burst / outbursts . sgr - like bursts did not produce these type of variations in broad - band red noise , however we believe that this is limited with the sensitivity of rxte ( also its background ) and statistical quality of rxte light curves . we calculate that 1e 2259 + 586 displays high level of broad - band noise lasting for a timescale of 2 years following its outburst and glitch . in addition , we observed significant rise of the rms broad - band noise level in 1e 1048.1 - 5937 for about 1.95 years , after the burst epoch at 53185 mjd . we believe that this noise can relate to the existence of remnant disks suggested to exist around these objects , but the timescale of changes after burst / outbursts are inconsistent with the observations in x - ray binaries . we stress that the changes in the broad - band noise levels are related to resonant cyclotron up - scattered radiation in axp and sgr spectra . correlation with long bursts and outbursts indicates an origin in the variability associated with the magnetar coronae . the typical timescales of existence of magnetar coronae of 1 - 10 years is consistent with the long term variation of bln detected in our work . we thank an anonymous referee for critical reading of the manuscript which improved this paper greatly . we , also , thank altan baykal and ali alpar for helpful comments . bk acknowledges support from tbitak , the scientific and technological research council of turkey , through project 106t040 . sb acknowledges an esa fellowship and support from the turkish academy of sciences with tuba - gebip ( young distinguished scientist award ) fellowship .

we present the broad - band noise structure of selected anomalous x - ray pulsars ( axps ) and soft gamma repeaters ( sgrs ) in the 2 - 60 kev energy band . we have analyzed rossi x - ray timing explorer proportional counter array archival light curves for four axps and one sgr . we detect that the persistent emission of these sources show band limited noise at low frequencies in the range 0.005 - 0.05 hz varying from 2.5@xmath0 to 70@xmath0 integrated rms in times of prolonged quiescence and following outbursts . we discovered band - limited red noise in 1e 2259 + 586 only for @xmath12 years after its major 2002 outburst . the system shows no broad - band noise otherwise . although this rise in noise in 1e 2259 + 586 occurred following an outburst which included a rotational glitch , the other glitching axps showed no obvious change in broad - band noise , thus it does not seem that this noise is correlated with glitches . the only source that showed significant variation in broad - band noise was 1e 1048.1 - 5937 , where the noise gradually rose for 1.95 years at a rate of @xmath13.6@xmath0 per year . for this source the increases in broad - band noise was not correlated with the large increases in persistent and pulsed flux , or its two short sgr - like bursts . this rise in noise did commence after a long burst , however given the sparsity of this event , and the possibility that similar bursts went unnoticed the trigger for the rise is noise in 1e 1048.1 - 5937 is not as clear as for 1e 2259 + 586 . the other three sources indicate a persistent band - limited noise at low levels in comparison . [ firstpage ] stars : neutron stars : oscillations stars : coronae x - ray : stars .

this paper presents an analysis of the @xmath0-carbon cross section data from the karmen @xcite and lsnd @xcite experiments , within the context of electron neutrino oscillations at high @xmath1 . in a two - neutrino oscillation formalism , the probability for @xmath0 disappearance is given by : @xmath4 where @xmath5 is the mixing angle ; @xmath6 , in ev@xmath7 , is the difference between the squared neutrino masses ; @xmath8 , in m , is the distance from the neutrino source to the detector ; and @xmath9 , in mev , is the neutrino energy . this analysis exploits the fact that karmen and lsnd were located at @xmath1017.7 m and 29.8 m respectively . we use the consistency between the cross section measurements to place strong constraints on @xmath0 disappearance at @xmath11 ev@xmath7 . this study is motivated by recent results that can be interpreted as oscillations with @xmath12 ev@xmath7 . the strongest evidence comes from the lsnd experiment , which observed a @xmath13 signal corresponding to an oscillation probability of @xmath14 @xcite . recent miniboone antineutrino data @xcite are in agreement with lsnd , although with less significance , while the miniboone neutrino data do not support @xmath15 oscillations @xcite . high @xmath1 muon - to - electron flavor appearance can not be explained in a three neutrino mixing model that also incorporates `` solar '' and `` atmospheric '' oscillations @xcite . as a result , these data have inspired models with three active and one sterile ( 3 + 1 ) or 2 sterile ( 3 + 2 ) neutrinos . sterile neutrinos ( @xmath16 ) do not interact via the weak interaction , but can mix with and cause oscillations between , the active flavors . these models predict a @xmath17 signal @xcite with a large @xmath1 ( on the order of a few ev@xmath7 ) compared to the splittings between the light states ( of order @xmath18 and @xmath19 ev@xmath7 ) . therefore , one can take the three light states to be effectively degenerate . this degeneracy simplifies the 3 + 1 model to an approximate two - neutrino oscillation model for both appearance and disappearance . as a result , eq . [ osc ] will be applicable to the following discussion , where we will use @xmath20 as the mixing angle relevant to @xmath0 disappearance . recently , a reanalysis of reactor @xmath21 flux predictions @xcite has provoked further interest in electron flavor disappearance in 3 + 1 models @xcite . this new analysis resulted in a shift of the ratio of reactor data - to - prediction from @xmath22 to @xmath23 . this deficit with respect to prediction is called the `` reactor anomaly '' in this paper . this can be taken as indication of @xmath24 in a 3 + 1 model at 98.6% cl @xcite . the best fit is @xmath25 ev@xmath7 and @xmath26 @xcite . indications of @xmath0 disappearance have arisen from calibration data taken by the sage @xcite and gallex @xcite experiments . these used megacurie sources of @xmath27cr and @xmath28ar to calibrate the @xmath29 experiments . the data from sage and gallex are consistent , and show a measured - to - predicted ratio of @xmath30 @xcite . we refer to this as the `` gallium data '' in this paper . this can be interpreted as a 2.7@xmath3 indication of @xmath17 oscillations @xcite . the best fit in a 3 + 1 model corresponds to a @xmath31 ev@xmath7 and @xmath32 @xcite . this apparent @xmath0 disappearance signal leads to the argument @xcite that @xmath17 , consistent with the gallium data , must be applied to the intrinsic @xmath0 background of the miniboone @xmath15 search @xcite . , title="fig:",width=264 ] the results of the 3 + 1 models have provoked substantial interest within the community , but they are not decisive . as a result they are a prime motivation for further studies of electron neutrino disappearance @xcite . there are few opportunities for precision @xmath0 disappearance searches , since most beams have large uncertainties in the normalization and energy distribution for the @xmath0 and @xmath33 beam contents . however , decay - at - rest ( dar ) neutrino beams can provide a unique window on electron - flavor neutrino oscillations . karmen and lsnd were dar experiments that ran in the 1990 s using 800 mev protons on target . the isotropic dar flux , shown in fig . [ flux ] , has equal @xmath34 , @xmath35 and @xmath0 content with a well - understood energy spectrum described by weak decay physics . the beam energy extends to 52.8 mev . the normalization is known to 10% , with the uncertainty dominated by the pion production rate per incident proton @xcite . karmen ran at the isis facility at rutherford laboratory , with 200 @xmath36a of protons impinging on a copper , tantalum , or uranium target . the center of the nearly cubic detector was located at 17.7 m from the proton target , at an angle of 100@xmath37 . the liquid scintillator target volume was 56 m@xmath38 and consisted of 512 optically independent modules ( 17.4 cm @xmath39 17.8 cm @xmath39 353 cm ) wrapped in gadolinium - doped paper . more details are available in ref . @xcite . lsnd used protons from the lampf accelerator at los alamos national laboratory ( lanl ) , where a 1 ma beam of protons impinged on a water target . the center of the 8.75 m long , nearly cylindrical detector was located at 29.8 m from the target , at an angle of 12@xmath37 from the proton beam direction . this was an unsegmented detector with a fiducial mass of 167 tons of oil ( ch@xmath40 ) , lightly doped with b - pbd scintillator . more details are available in ref . @xcite . both experiments measured @xmath41 scattering . in this two - body interaction , with @xmath42-value of 17.3 mev , the neutrino energy can be reconstructed by measuring the outgoing visible energy of the electron . the @xmath43n ground state is identified by the subsequent @xmath44 decay , @xmath45 , which has a @xmath42-value of 16.3 mev and a lifetime of 15.9 ms . the karmen and lsnd cross section measurements for @xmath46 @xcite , in energy bins , are compared in fig . [ xsecs ] . the corresponding flux - averaged cross sections measured by karmen and lsnd are given in table [ fluxavetab ] . for completeness , we also list the flux - averaged cross section for the lanl e225 experiment @xcite , which was located 9 m from a dar source . ( e225 did not publish energy - binned cross section measurements . ) the agreement between all three experiments is excellent . the measured cross sections are compared to predictions by fukugita , _ et al._@xcite and by kolbe _ et al._@xcite in fig . [ xsecs ] . both models follow an @xmath47 form , where @xmath48 mev . this energy dependence arises because the interaction is an allowed transition , converting the @xmath49 ( @xmath43c ) state to the @xmath50 ( @xmath43n ) state . the fukugita prediction is calculated within the `` elementary particle model '' ( ept ) and has an associated 12% normalization uncertainty @xcite . other ept predictions , include donnelly @xcite and mintz , _ et al . _ , are given in table [ fluxavetab ] . for comparison , the kolbe , _ et al . _ , calculation @xcite is performed within a `` continuum random phase approximation '' ( crpa ) approach . a discussion of the relative merits of ept versus crpa models for describing this process appears in ref . @xcite . from a strictly experimental point of view , both ept and crpa models fit the data well . to be clear , these theoretical results are true predictions rather than fits , since they were published well before the karmen and lsnd results . compared to the theoretical prediction of fukugita , _ et al . _ ( solid line ) , based on the ept model , and kolbe , _ et al . _ ( dashed line ) , based on the crpa model . [ xsecs ] , title="fig:",width=283 ] [ c]l|c|cexperiment ( dist . ) & flux - averaged cross section & ref . + karmen ( 17.7 m ) & @xmath51 cm@xmath7 & @xcite + lsnd ( 29.8 m ) & @xmath53 cm@xmath7 & @xcite + e225 ( 9 .0 m ) & @xmath54 cm@xmath7 & @xcite + prediction & flux - averaged cross section & ref . + fukugita _ et al . _ & @xmath55 cm@xmath7 & @xcite + mintz _ et al . _ & @xmath56 cm@xmath7 & @xcite + donnelly & @xmath57 cm@xmath7 & @xcite + kolbe _ et al . _ & @xmath58 cm@xmath7 & @xcite + the allowed regions for @xmath17 oscillations are determined from a comparison of the lsnd and karmen data with respect to the fukugita prediction . for a given oscillation hypothesis ( @xmath1 and @xmath59 ) , we calculate a combined @xmath60 for lsnd and karmen with respect to the prediction using the statistical error for each data point and employing three pull terms as a method to incorporate systematic uncertainties . the first pull term represents the correlated normalization error . as noted in the karmen paper @xcite , lsnd and karmen have a 7% systematic error on the neutrino flux normalization from the flux simulation that is correlated between the two experiments @xcite . this is combined in quadrature with the 12% systematic error on the normalization for the fukugita prediction to give the correlated normalization pull term in the @xmath60 calculation . the remaining uncorrelated normalization uncertainties for each experiment are 7% for lsnd @xcite and 5% for karmen these uncertainties are used as the two other pull terms in the @xmath60 calculation . to determine the 90% cl allowed regions in @xmath1 and @xmath59 , we marginalize over the three normalization pull parameters and use a @xmath61 requirement for the two - degrees - of - freedom excluded region . , title="fig:",width=264 ] [ c]lcrr & & + fit type & @xmath1(ev@xmath7 ) & & + best fit & 7.49 @xmath62 0.39 & & + no osc & & & + gallium & 2.24 ( fixed ) & & + reactor anomaly & 1.78 ( fixed ) & & + & & & + & & + fit type & @xmath1(ev@xmath7 ) & & + best fit & 7.49 @xmath62 0.39 & & + no osc & & & + gallium & 2.24 ( fixed ) & & + reactor anomaly & 1.78 ( fixed ) & & + the results of the fits using the fukugita prediction are shown in fig . [ fits ] . table [ chi2 ] reports the @xmath60 and degrees of freedom ( dof ) for various joint fits to the lsnd and karmen data points . the fit without oscillations ( no osc ) , shown as the long - dashed line in fig . [ fits ] , has a @xmath63 probability of 91.5% and is only excluded at the 1.7@xmath3 level . as a result , we use the data to set a limit on @xmath0 disappearance and calculate the 95% cl exclusion region shown in fig . [ contour ] . the best fit , indicated by the solid lines in fig . [ fits ] , is at @xmath64 ev@xmath7 and @xmath65 . comparing the data to an oscillation model with the best - fit gallium parameters illustrates the strong disagreement , though we note that the gallium fit had a rather shallow minimum @xcite . the gallium fit reported in table [ chi2 ] and shown as the dashed line on fig . [ fits ] is poor . this point has a @xmath60 probability of less than @xmath66 and is , therefore , ruled out at @xmath67 . ( the @xmath63 for this point has a probability of @xmath68 , which corresponds to a @xmath69 exclusion . ) most of the gallium allowed region , indicated at 68% and 90% cl on fig . [ contour ] , is excluded at 95% cl by this analysis . disappearance limit from the fukugita ( ept ) fit ( solid , black line ) compared to the predicted sensitivity ( dotted line ) . also shown is the 68% ( darker , shaded region ) and 90% ( lighter , shaded region ) contours from the gallium experiments . the dashed line is the kolbe ( crpa ) fit . [ contour ] , title="fig:",width=283 ] + as discussed above , all models tend to follow a @xmath47 dependence . nevertheless , small differences between the fukugita ( ept ) and kolbe ( crpa ) predictions , shown in fig . [ xsecs ] , allow a test for model dependence . the kolbe fit proceeds in the same way as for the fukugita model . the resulting @xmath60 values for the fits are given in table [ chi2 ] . the comparisons of the fits with the data are indistinguishable from those shown in fig . [ fits ] and so are not shown here . this leads to the conclusion that there is no substantial systematic effect from the energy dependence of the underlying cross section model . the 95% cl exclusion limit from the kolbe fit is also shown in fig . [ contour ] as the dashed contour and is very similar to the fukugita contour . the excluded region in fig . [ contour ] is better than the expected sensitivity region ( dotted contour ) calculated for an underlying null oscillation hypothesis . as a way of quantifying this difference , using the fukugita model fit , the gallium data point is ruled out in a @xmath60 analysis at 3.6@xmath3 , while the sensitivity would have predicted that , for an average experiment with no signal , the gallium point would be ruled out at 2.8@xmath3 . this strong limit with respect to the sensitivity is not unlikely ; we find that 11% of simulated experiments have a high @xmath1 95% cl limit for @xmath70 at or below the 0.34 95% cl limit of this analysis . these studies also show that @xmath63 is a good statistic for determining the exclusion regions since 10% of the simulated experiments have a @xmath63 value for the null oscillation hypothesis greater than 4.61 as expected for two degrees of freedom . disappearance limit from the combined fit ( solid , black line ) compared to individual fits to karmen data ( dashed ) and lsnd data ( dotted ) . this is overlaid on the gallium 90% and 68% cl allowed regions . all fits use the fukugita model . [ individual ] , title="fig:",width=283 ] + the combined fit to karmen and lsnd is stronger than fits to the the individual data sets because the detectors are at different distances . requiring the proper @xmath8 as well as @xmath9 dependence adds an important constraint to the oscillation fit . the results of the individual fits are compared to the combined fit in fig . [ individual ] . a fit to only the karmen data yields a best fit @xmath60 of 2.46 for 4 degrees of freedom , with the parameters @xmath71 and @xmath72 ev@xmath7 . the @xmath60 for the null fit was 7.05 for 6 degrees of freedom . a fit to only the lsnd data results in a best fit @xmath60 of 2.27 for 3 degrees of freedom , with the parameters @xmath73 and @xmath74 ev@xmath7 . the @xmath60 for the null fit was 3.29 for 5 degrees of freedom . considering fits to these two data sets separately allows interpretation of certain features in the combined fit . we see in fig . [ individual ] that the karmen data dominates at high @xmath1 because of statistics and that the large variations in the limit are driven by this data set . these variations appear because the event energy distribution spans a limited range . as a result , there are oscillation parameters for which the karmen distance allows the experiment to either be very sensitive or very insensitive to disappearance . the lsnd data set allows full oscillation at @xmath1 of 3.9 ev@xmath7 , but karmen data does not . at low @xmath1 , the sensitivity is dominated by the lsnd data although the combination with karmen is significantly better . the limit presented here and the gallium data represent the only electron neutrino disappearance results in this @xmath1 range . comparison to other data sets require interpretation within models . specific global analyses are beyond the scope of this paper , however , we can consider the impact of the limit , in general . cpt conservation requires that @xmath0 and @xmath21 disappearance should occur at the same rate . because cpt conservation is embedded in all field theories , comparison of electron flavor neutrino and antineutrino disappearance is widely regarded as interesting . the relavent antineutrino disappearance data comes from the reactor experiments . the recently published reactor anomaly hints at oscillations , and the parameters from the @xmath21 disappearance data @xcite can be compared to the @xmath0 results assuming cpt conserving models . in fig . [ reactcontour ] , we overlay the reactor anomaly allowed region at 68% and 90% cl with the 95% cl limit from this paper using fukugita . in this figure , we use a log scale for @xmath75 so that the reactor allowed region is clear . one can see that portions of the allowed reactor space will be excluded in fits that require cpt conservation . however , the reactor anomaly best fit parameters give a @xmath63 probability of 90.8% , which lies outside of the 95% cl excluded region from the cross section analysis limit , and one expects portions of the allowed region to survive in a global fit . the comparison of the best fit with data , shown as the thin line with dots in fig . [ fits ] , is reasonably good and similar to the no oscillation " case , illustrating why the karmen / lsnd data does not speak to the entire region of the reactor anomaly . this is consistent with good the @xmath60/dof for the best fit point of the reactor anomaly , reported in table [ chi2 ] . disappearance limit from the fukugita ( ept ) fit ( solid , black line ) and kolbe ( crpa ) fit compared to the 90% ( shaded region ) contours from the reactor anomaly . [ reactcontour ] , title="fig:",width=283 ] + comparisons to electron - to - muon flavor appearance data is much more model dependent than the comparison of @xmath21 to @xmath0 disappearance . in this case , one must go beyond choosing whether or not to assume cpt conservation , to include the number of sterile neutrinos ( typically , one assumes 3 + 1 or 3 + 2 models ) ; the mass heirarchy ( a 1 + 3 + 1 model has recently been published @xcite as a variation of the two sterile neutrino case ) ; and whether or not to include cp violation . inclusion of the appearance data also , necessarily , requires reference to muon - flavor disappearance as well as electron flavor disappearance , since these are all realted . as an example , in a 3 + 1 model : @xmath76 where @xmath77 represents the element in the 4-neutrino mixing matrix . the disappearance limit we present here , when interpreted within eq . [ ediseq ] , substantially limits the range of @xmath78 , and will considerably reduce the space of allowed values for matrix elements that can describe muon - to - electron flavor appearance in neutrino mode , as may be implied by the miniboone low energy excess @xcite . while the correspondences between experimentally measured mixing angles and the underlying matrix elements for two - sterile - neutrino models is more complicated , the basic point still applies this analysis has used the @xmath79 cross section data from lsnd and karmen to constrain the amount of @xmath0 disappearance oscillations at high @xmath1 . the good agreement between the data sets and with the theory , despite different distances of detectors from the source , leads to an 95% cl exclusion region which extends down to @xmath80 at high @xmath1 and to considerably lower values for some @xmath1 ranges . comparison to another underlying cross section model does not significantly change the excluded region . the data are in disagreement with the only other @xmath0 disappearance data set in this @xmath1 region , which comes from the gallium experiments . large portions of the allowed region for @xmath0 disappearance analysis of the gallium calibration data are ruled out at @xmath295% cl . as a benchmark , the best fit point for gallium is excluded at 3.6@xmath3 . this new limit also severely restricts models addressing the miniboone results , such as ref . @xcite , that incorporate @xmath0 disappearance to explain the observed energy distribution . assuming cpt conservation , this data set can be compared to the @xmath33 disappearance from reactor data . the reactor anomaly best fit point in this case is within the allowed 95% cl contour but some regions of the allowed region in ref . @xcite are excluded . l. b. auerbach _ et al . _ [ lsnd collaboration ] , phys . c * 64 * , 065501 ( 2001 ) . a. aguilar _ et al . _ [ lsnd collaboration ] , phys . d * 64 * , 112007 ( 2001 ) . a. a. aguilar - arevalo _ et al . _ [ the miniboone collaboration ] , phys . lett . * 105 * , 181801 ( 2010 ) . a. a. aguilar - arevalo _ et al . _ [ the miniboone collaboration ] , phys . lett . * 98 * , 231801 ( 2007 ) . m. sorel , j. m. conrad and m. shaevitz , phys . d * 70 * , 073004 ( 2004 ) . g. karagiorgi , z. djurcic , j. m. conrad , m. h. shaevitz and m. sorel , phys . d * 80 * , 073001 ( 2009 ) [ erratum - ibid . d * 81 * , 039902 ( 2010 ) ] . t. a. mueller _ et al . _ , phys . rev . c * 83 * , 054615 ( 2011 ) .

this paper presents a combined analysis of the karmen and lsnd @xmath0-carbon cross section measurements within the context of a search for @xmath0 disappearance at high @xmath1 . karmen and lsnd were located at 17.7 m and 29.8 m respectively from the neutrino source , so , the consistency of the two measurements , as a function of antineutrino energy , sets strong limits on neutrino oscillations . most of the allowed region from the @xmath0 disappearance analysis of the gallium calibration data is excluded at @xmath295% cl and the best fit point is excluded at 3.6@xmath3 . assuming cpt conservation , comparisons are also made to the oscillation analyses of reactor antineutrino data .

over the past five years , a major controversy has developed about the high - pressure melting curves of transition metals . according to static compression experiments , performed using diamond anvil cells ( dac ) at pressures from ambient up to @xmath3 gpa ( 1 mbar ) , the melting temperatures @xmath4 of many transition metals , particularly those having the b.c.c . crystal structure , change by no more than a few hundred k over this pressure range @xcite . in some cases , this finding seems to be in gross conflict with shock experiments , which indicate an increase of @xmath4 of several thousand k over the same pressure range @xcite . theoretical work @xcite based directly or indirectly on density functional theory ( dft ) generally supports the shock data . we report here a detailed dft study of the high - pressure melting curve of mo , a metal for which there are some of the largest differences between dac measurements and other data ( see e.g. fig . 2 in ref . @xcite ) . there has already been quite extensive theoretical work on the high-@xmath5 melting of metals . some of this has been motivated by the desire to understand the properties of solid and liquid fe in the earth s core @xcite . dft - based calculations of the fe melting curve @xcite up the pressure at the boundary between the solid inner core and the liquid outer core provide one of the important ways of constraining the temperature distribution in the core . disagreements with dac measurements on fe and other transition metals are cause for concern , because if dft were shown to be seriously in error , the reliability of dft for the study of planetary interiors would be called into question . but even without this practical motivation , a major disagreement between dft predictions and experimental data must be taken seriously , because it suggests an unexpected failure either of commonly used dft approximations , or of apparently well established experimental techniques . reassuringly , dft melting curves agree very closely with experiment for some metals , including al @xcite and cu @xcite . the large disagreements arise mainly for transition metals , and the suggestion is that they are linked to d - band bonding . since the melting slope @xmath2 is equal by the clausius - clapeyron relation to @xmath6 , where @xmath7 and @xmath8 are the volume and entropy of fusion , and since @xmath8 is unlikely to have exceptionally large values , a very low @xmath2 is likely to be due to a low volume of fusion . it has been argued @xcite that a low @xmath7 might be expected for b.c.c . transition metals , because the liquid may be more close packed than the solid . but recently , an additional argument has been advanced for low @xmath2 values @xcite . this argument depends on the fact that the distribution of conduction electrons between s - p states and d states is known to depend on the degree of compression and on the crystal stucture @xcite . the suggestion is that the ( assumed ) change of coordination on going from b.c.c . solid to liquid leads to a change of electronic structure , and hence a change in the electron distribution between s - p and d states , and that this redistribution stabilises the liquid and lowers @xmath4 . one of the important purposes of the present work is to use dft molecular dynamics ( m.d . ) simulations to test these suggestions for the case of mo . there has been previous dft work @xcite on the high - pressure melting of mo . the early work of moriarty @xcite employed a many - body total energy function derived from first - principles `` generalized pseudopotential theory '' @xcite . this approximates the total energy of the system in terms of volume dependent 1- , 2- , 3- and 4-body interatomic potentials , and accounts for the angular forces that are known to be important in transition metals . this form of total - energy function is designed to be fully transferable between different structures , and should be valid for both the solid and the liquid state . the model total - energy function was used in m.d . simulations to determine the melting curve by two methods . the first method consisted of cycling the simulated system up and down through the melting point at fixed volume . the results were cross - checked against a second method based on the calculation of the free energies of the solid and the liquid . this was pioneering work , but its quantitative accuracy can be questioned , because some of the phonon frequencies predicted by the total energy function agreed rather poorly with experiment . in addition , the surprising claim was made that @xmath4 could be changed by up to a factor of two by the inclusion of thermal electronic excitations , which were treated only crudely . the predicted melting curve was consistent with shock data , but was far above the curve given by recent dac measurements : the difference of @xmath4 values amounts to @xmath9 k at @xmath10 gpa . much more recently , direct dft m.d . simulation has been used by belonoshko _ et al . _ @xcite to map out the mo melting curve up to @xmath11 gpa . their melting curve , like that of moriarty @xcite , is consistent with shock data , but disagrees strongly with dac data . however , even though the simulations employed an implementation of dft that is expected to be accurate , the method used to map the melting curve may still give inaccurate results , as noted by the authors . their method was to heat the simulated solid at constant volume until the internal energy , pressure , radial distribution function and self - diffusion coefficient showed discontinuities attributable to melting . this approach gives an upper bound to @xmath4 , but may overestimate it significantly , because of superheating . the authors estimated that the error in @xmath4 due to superheating should be @xmath12 % , but the arguments for this estimate are indirect . the present new work on mo melting has several aims . the first aim is to calculate more accurately the melting curve that follows from the adopted dft exchange - correlation functional @xmath13 . ( we present tests showing that the gga - pbe functional is a good choice . ) it is only by doing this that the possible reasons for the disagreement between theory and experiment can be narrowed down . we have shown in our recent work on other materials that the errors in computed dft @xmath4 values can be reduced to a few percent , and we have described several techniques for doing this @xcite . the technique used here is the `` reference coexistence method '' @xcite . this requires the accurate fitting of an empirical `` reference '' total energy function to dft m.d . simulations on the solid and the liquid ; the melting curve of the reference model is then calculated from simulations on large systems consisting of coexisting solid and liquid ; as an essential last step , free - energy corrections for the difference between the reference and dft total - energy functions are used to correct the melting curve . this technique was successfully used in our work on the melting of cu @xcite , and it has been shown to give results in excellent agreement with an alternative technique , in which dft free energies of the solid and liquid are calculated @xcite . since it is a ` thermodynamic ' technique relying on equality of gibbs free energies of the two phases , it can not suffer from superheating problems . a second important aim is to study the differences of atomic and electronic structure of the coexisting solid and liquid , to provide an improved understanding of the factors that determine the melting curve of mo . the empirical reference model used to determine the dft melting curve has the form of the embedded atom model ( eam ) @xcite . a useful side benefit of the work is that we obtained a parameterized eam that mimics quite well the dft total - energy function of high-@xmath1 solid and liquid mo . this model reveals important features of the energetics of mo at high-@xmath5 and high-@xmath1 , and we expect it to be useful in future modelling work on this metal . the remainder of the paper is organized as follows . a brief summary of the dft modelling techniques is given in sec . [ sec : techniques ] , followed by an outline of the reference coexistence technique . then , sec . [ sec : tests ] details the rather extensive tests we have performed to ensure that the techniques deliver an accurate description of the energetics and the vibrational and electronic properties of mo . [ sec : melting ] presents our results for the dft melting curve of mo up to 400 gpa , together with the entropy and volume of fusion as a function of pressure ; our comparison of the atomic and electronic structures of the high-@xmath1 b.c.c . solid and the liquid is reported at the end of the section . a discussion of all the results and their relation with previous work is given in sec . [ sec : discussion ] , followed by our conclusions . a comprehensive description of dft methods as applied to the modelling of condensed matter is given in recent books @xcite . as is well known , there is one and only one uncontrollable approximation in dft , namely the approximation used for the exchange - correlation functional @xmath13 . there is abundant evidence that commonly used @xmath13 functionals yield accurate results for a range of properties of transition - metal crystals , including equilibrium lattice parameter , elastic constants , phonon frequencies , @xmath14 equation of state , solid - state phase boundaries , etc @xcite . dft provides the basis for our present understanding of the electronic structure of transition - metal crystals as a function of pressure . there is also some evidence that it accurately reproduces the structure of liquid transition metals @xcite . nevertheless , we considered it necessary to test the accuracy of different @xmath13 in the case of mo , as we report in sec . [ sec : tests ] . a completely separate issue from the choice of @xmath13 is the _ implementation _ of dft that is used . this concerns mainly the way that the electron orbitals are represented . for simulations of the high-@xmath1 solid and the liquid , dft molecular dynamics ( m.d . ) must be used , and for this we use the paw ( projector augmented wave ) technique @xcite , which is generally regarded as the most accurate for m.d . the paw method developed by blchl @xcite efficiently combines some of the features originally devised within both linear augmented - plane - wave ( lapw , briefly described below ) and pseudopotential approaches . in paw , a linear transformation between the all - electron and pseudized wavefunctions is defined in terms of all - electron and pseudized partial waves and a set of projector functions localized on the atoms . there is a simple formal relationship between paw and the ultra - soft pseudopotential method of vanderbilt @xcite , and computationally the two approaches are almost equivalent @xcite . for m.d . , we use the method known as born - oppenheimer dynamics , in which the self - consistent ground state is recalculated at each new m.d . time - step . given the criticisms that are sometimes made of dft work on melting properties @xcite , is important to be clear that this means the recalculation of the electronic structure of the entire simulated system , so that changes of electronic structure on going from solid to liquid at any degree of compression are fully included . although paw is well suited to m.d . simulation , we regard it as essential to check its accuracy against still more accurate implementations of dft , and to do this we have compared with the predictions of the fp - lapw ( full - potential linearized augmented plane wave ) technique , which was developed for high - precision calculations on crystals @xcite . in this approach , space is divided into spherical regions centred on the ions , and the interstitial space between the spheres . accurate solution of schrdinger s equation on a radial grid for each angular momentum within each sphere allows both valence and core states to be treated to arbitrarily high precision . the fp - lapw results to be presented include fully relativistic treatment of the core states and scalar relativistic treatment of the states in valence @xcite . in fp - lapw calculations on crystalline mo , we can ensure that they are fully converged with respect to all technical parameters , so that all errors of _ implementation _ are negligible , and the only approximation is that due to @xmath13 itself . all our paw and fp - lapw calculations were performed using the vasp code @xcite and the wien2k code @xcite , respectively . normally , the aim of dft calculations is to obtain the electronic ground state for given ionic positions . however , because of the high @xmath1 involved , it is essential in the present work to include thermal electronic excitations , so that for any given ionic positions we must determine the orbitals and occupation numbers that self - consistently minimize the electronic _ free _ energy , using the @xmath15 version of dft originally developed by mermin @xcite . we know this is essential , because work on other transition metals @xcite shows that the electronic specific heat becomes comparable with the vibrational contribution for @xmath16 k. without this , the free energy difference between solid and liquid might be seriously in error . all our m.d . simulations are done in the canonical @xmath17 ensemble , with the electronic @xmath1 set equal to the @xmath1 of the ensemble . values of technical parameters ( @xmath18-point sampling , plane - wave cut - off , etc . ) will be given when we present the calculations . there are three steps in the reference coexistence technique @xcite . first , an empirical reference model is fitted to _ ab initio _ simulations of the solid and the liquid at thermodynamic states close to the expected melting curve . then , the reference model is used to perform simulations on large systems in which solid and liquid coexist , so as to obtain points on the melting curve of the model . finally , the differences between the reference and _ ab initio _ total energy functions are used to correct the melting properties of the fitted model so as to obtain the _ ab initio _ melting properties . in this work , we have used the embedded - atom model ( eam ) @xcite as the reference model . the total energy function @xmath19 of this model for a system of @xmath20 atoms has the form : @xmath21^{1/2 } .\ ] ] where @xmath22 is the separation of atoms @xmath23 and @xmath24 . the first term on the right represents an inverse - power repulsive pair potential , while the second ( embedding ) term describes the @xmath25-band bonding . the model is specified by the characteristic length @xmath26 , the energy scale @xmath27 , the dimensionless coefficient @xmath28 characterizing the strength of the embedding energy , and the embedding and repulsive exponents @xmath29 and @xmath30 . we wish @xmath31 to mimic as closely as possible the _ ab initio _ total energy function @xmath32 , and we achieve this by adjusting the eam parameters . the method for fitting @xmath19 to @xmath33 is designed so as to minimize the corrections to the melting curve caused by the difference @xmath34 . we therefore recall the correction scheme before describing the fitting itself . for given @xmath5 and @xmath1 , the difference @xmath35 between the gibbs free energies of the _ ab initio _ liquid and solid deviates from the corresponding difference @xmath36 of the reference liquid and solid , and we write : @xmath37 it is the shift @xmath38 caused by changing the total - energy function from @xmath19 to @xmath33 that causes the shift of melting temperature at given pressure . to first order , the latter shift is @xcite : @xmath39 where @xmath40 is the difference between the entropies of liquid and solid ( i.e. the entropy of fusion ) of the reference system , and @xmath41 is @xmath42 evaluated at the melting temperature of the reference system . the shift @xmath42 is the difference of shifts of gibbs free energies of liquid and solid caused by the shift @xmath34 of total energy function . we find it convenient to perform our simulations at constant volume and temperature . under these conditions , the shift of helmholtz free energy @xmath43 arising from @xmath44 is given by the well - known expansion : @xmath45 where @xmath46 , @xmath47 , and the averages are taken in the reference ensemble . from @xmath43 , we obtain the shift of gibbs free energy at constant pressure as : @xmath48 where @xmath49 is the isothermal compressibility and @xmath50 is the change of pressure when @xmath19 is replaced by @xmath51 at constant @xmath52 and @xmath1 . in fitting the eam @xmath19 to the _ ab initio _ @xmath33 so as to minimize the corrections we have just described , we see from eq . ( [ eq : helmholtz_exp ] ) and ( [ eq : isocorisobar ] ) that the effects of @xmath53 , @xmath54 and @xmath50 all need to be made small . concerning @xmath55 , we note that addition of a position - independent constant to either @xmath33 or @xmath19 has no effect on the properties of the solid or the liquid , since it simply redefines the energy zero , so that a large value of @xmath53 is not in itself significant . however , it is crucially important that @xmath53 should have almost the same value in the solid and the liquid , because otherwise there would be a large shift of @xmath42 . we also seek to make @xmath54 and @xmath50 small in both phases . to satisfy all these requirements , we perform long _ ab initio _ m.d . simulations of the solid and the liquid at thermodynamic states near the expected melting curve . a large number of statistically independent configurations are then drawn from both these runs , so that we get a large set of @xmath33 values for a collection of configurations representative of the solid and the liquid . our procedure is then to minimize the mean square fluctuations of @xmath44 over this whole set : with @xmath56 the mean value of @xmath44 over the whole set , we minimize @xmath57 , where @xmath58 . note that this @xmath57 is not the same quantity as the @xmath54 appearing in eq . ( [ eq : helmholtz_exp ] ) , because @xmath57 characterizes the fluctuations of @xmath44 over a set of configurations drawn from both liquid and solid . minimization of @xmath57 has the effect of reducing simultaneously the difference of @xmath53 between solid and liquid , and of reducing @xmath54 in the two phases . in order to reduce also @xmath59 in the two phases , we add @xmath59 into the quantity to be minimized , with a suitable weight . as will be described in sec . [ sec : results ] , we find it necessary to refit the eam in this way for different ranges of pressure along the melting curve . a convenient way of characterizing the quality of fit of @xmath19 to @xmath33 is obtained by dividing @xmath57 by the mean square value of the fluctuation of _ ab initio _ energy @xmath60 , where @xmath61 is the _ ab initio _ energy averaged over the collection of solid and liquid configurations . the dimensionless quantity characterizing the fit is then @xmath62^{1/2}$ ] ; the smaller the value of @xmath63 , the better is the fit . following our previous work @xcite , our simulations of coexisting solid and liquid with the reference model are performed in the @xmath64 ensemble . starting with a supercell containing the perfect b.c.c . crystal , we thermalize it at a temperature slightly below the expected melting temperature . the system remains in the solid state . the simulation is then halted , and the positions of the atoms in one half of the supercell are held fixed , while the other half is heated to a very high temperature ( typically eight times the expected melting temperature ) , so that it melts completely . with the fixed atoms still fixed , the molten part is then rethermalized to the expected melting temperature . finally , the fixed atoms are released , thermal velocities are assigned , and the whole system is allowed to evolve freely at constant ( @xmath65 ) for a long time ( normally more than 60 ps ) , so that the solid and liquid come into equilibrium . the system is monitored by calculating the average number of particles in slices of the cell taken parallel to the boundary between the solid and liquid . with this protocol , there is a certain amount of trial and error to find the overall volume that yields the coexisting solid and liquid system . ( an example of the density profile obtained when the two phases are in stable coexistence will be given in sec . [ sec : melting ] . ) our main simulations were done on cells containing 6750 atoms , constructed as a @xmath66 b.c.c . supercell , the long axis being perpendicular to the initial liquid - solid boundary . we tested the adequacy of this system size by repeating the simulations with larger systems ( around 10000 atoms ) , and found no change in the results . finally , our best value of the _ ab initio _ melting temperature is obtained by adding to the reference melting temperature the correction @xmath67 given by eq . ( [ eq : correctiont ] ) . the reference entropy of fusion @xmath40 needed in this equation is obtained by performing independent reference m.d . simulations of the solid and the liquid in the @xmath17 ensemble at the temperatures at which we made the coexistence simulations , using at each of these temperatures the solid and liquid volumes that yield the coexistence pressure . these simulations give the enthalpy of fusion @xmath68 , from which we obtain @xmath40 from the relation @xmath69 . to obtain the values of @xmath41 needed in eq . ( [ eq : correctiont ] ) , we again perform reference m.d . simulations of solid and liquid at the coexistence temperatures and volumes , and calculate @xmath33 for a statistically independent set of configurations drawn from these simulations . these @xmath33 values must , of course , be fully converged with respect to system size and @xmath70-point sampling . we find that for a system of 125 atoms and with @xmath33 calculated with a @xmath71 @xmath70-point grid for both solid and liquid , @xmath33 is converged within @xmath72 mev / atom . these @xmath33 values are then used to compute @xmath53 and @xmath54 . the same calculations yield @xmath50 for solid and liquid , which we use in eq . ( [ eq : isocorisobar ] ) . to assess the accuracy of the techniques , we have carried out a number of tests on the zero - temperature crystal in the pressure range @xmath73 gpa . there are three important questions to analyze . first , we want to test the accuracy of different exchange - correlation functionals compared with experimental data . second , we aim to study the effect on paw results of including different electronic states in the valence set . third , fp - lapw calculations are performed to assess errors incurred by the paw approximation . in addition , we have done tests to demonstrate that paw correctly reproduces the changes of electronic structure with pressure given by the fp - lapw method . tests of phonon frequencies are also carried out , because of their relevance to the vibrational free energy of the system . at very high pressures , states that would normally be treated as core states may respond significantly to compression . in the case of mo , the @xmath74 states lie only @xmath75 ev below the fermi energy @xmath76 and we always include them in the valence set . the @xmath77 states lie considerably deeper at @xmath78 ev below @xmath76 , and we have examined the effect of including them . we show in fig . [ fig : vasp - lda - pbe ] paw results for the pressure as a function of volume at @xmath14 k , with and without the @xmath77 states included in the valence set , compared with experimental data @xcite . results are shown with both lda ( ceperley - alder parameterization @xcite ) and gga ( perdew - burke - ernzerhof form @xcite ) exchange - correlation functionals . all calculations were performed on a primitive b.c.c . cell , with a @xmath79 @xmath70-point grid and with energy cut - offs of @xmath80 ev ( without @xmath77 states ) and @xmath81 ev ( with @xmath77 states ) ; these settings ensure energy convergence to better than 1 mev / atom . with lda , inclusion of @xmath82 states makes a significant difference , and improves the agreement with experiment at high pressure , but with gga the effect of including @xmath82 states is very small . all the approximations deviate noticeably from experiment . at low pressures , lda underestimates the volume by about @xmath83 % , while gga overestimates it by @xmath84 % . at high pressures , @xmath85 gpa , lda volumes are only @xmath86 % below measured values , while gga continues to overestimate them by about @xmath87 % . to check the accuracy of paw itself , we have compared with fp - lapw calculations of the @xmath14 k @xmath88 relation , performed with the wien2k code @xcite . our fp - lapw calculations include fully relativistic effects of the core states and scalar relativistic treatment of the states in valence @xcite , and the tolerances are chosen so as to ensure energy convergence to better than 1 mev / atom @xcite . as shown in fig . [ fig : wien2kvasp ] , the fp - lapw @xmath88 results are almost indistinguishable from the paw results over the entire pressure range of interest , with both lda and gga functionals . this provides valuable confirmation of the accuracy of the paw technique , on which all our present calculations of melting properties are based . out of interest , we also performed fp - lapw calculations using the recently developed wu - cohen form of gga @xcite , which has been reported to give improved predictions of condensed - matter properties . this functional satisfies the same constraints used to construct the pbe functional , but the enhancement factor entering the exchange energy density is chosen to match closely the gradient expansion of svendsen and von barth @xcite for systems having a slowly varying density . for this reason , this functional is expected to perform well for solids but not so well for atoms or molecules , where the variation of the valence electron density is generally larger . we find that this functional is in much better agreement with the experimental @xmath88 data than either lda or gga(pbe ) . unfortunately , at the time most of the present work was done , we had no way of performing wu - cohen calculations within paw , so we were unable to do melting calculations with this new functional . the electronic structure of transition metals has been intensively studied , and changes of electronic structure with increasing pressure have been thoroughly investigated both experimentally and theoretically @xcite . the energetics of these metals is , of course , strongly dominated by the d - bands , but the s - p bands also play an important role . an effect that has been well recognised for many years is the pressure induced upward shift of the s - p band relative to the d band , caused by the greater spatial extent of the s - p orbitals @xcite . this relative shift means that increasing pressure causes a transfer of electrons from s - p to d bands , resulting in an increase of occupancy of the latter . it has been proposed recently @xcite that the large differences between theoretical predictions of melting curves and the results of static compression experiments for some transition metals may be due to an incorrect treatment of s - p @xmath89 d transfer in the dft - based simulations . in order to demonstrate that our paw calculations reproduce the electronic structure as given by the most accurate dft methods , we have made detailed comparisons of the electronic density of states ( dos ) calculated with paw and fp - lapw over the pressure range @xmath73 gpa . we show in fig . [ fig : doswien2kvasp ] the electronic dos from fp - lapw and paw calculated at 0 , 150 and 300 gpa , using the gga(pbe ) functional . we note the essentially perfect agreement between the two methods . at all pressures , the dos consists of two separate parts : a narrow peak at about 35 ev below the fermi energy , corresponding to 4p states , and a much broader distribution consisting of the multiple peaks due to the 4d bands superimposed on the slowly varying dos of the s - p bands . we note the characteristic feature of b.c.c . transition metals that the fermi energy falls in a deep minimum in the d - band dos . as expected , the widths of the 4p and 4d parts of the dos broaden markedly with increasing pressure . the pressure induced relative shift of s - p and d bands can not be clearly seen from the dos itself . however , it is very clear from the band structure , shown at pressures of 0 and @xmath90 gpa in fig . [ fig : bandstructure ] . the state lying @xmath91 ev below the fermi energy is the bottom of the s - p band , which lies well below the d states at zero pressure , but well above the lowest d states at @xmath92 gpa . we have checked that this relative shift is precisely reproduced by the paw calculations . this leaves little doubt that the paw techniques , on which all our melting calculations are based , correctly reproduce this important feature of the pressure dependent electronic structure . in sec . [ sec : melting ] , we will present results on the temperature dependence of the electronic dos in the solid and the liquid . the calculation of phonon frequencies is an important test of dft approximations , because of the detailed comparisons with experimental data that can be made . it is particularly important in the context of melting calculations , because , for the harmonic solid , errors in phonon frequencies translate directly into free energy errors , which are linked with errors in melting temperature . we present here our calculations of the phonon dispersion relations of mo at its experimental volume , using paw with gga(pbe ) exchange - correlation . the technique for calculating the phonon frequencies is the small displacement method , as implemented in our phon code @xcite , which we used in earlier work on fe , al and cu . in this method , the elements of the force - constant matrix are obtained by displacing atoms from the perfect - lattice positions and computing by dft the forces on all the atoms . for a b.c.c . crystal , it suffices to displace a single atom along the ( 111 ) direction . since all the calculations employ periodic boundary conditions , the displaced atom is at the centre of a periodically repeated supercell . to obtain accurate dispersion relations over the whole brillouin zone , this supercell must be large enough so that the elements of the force - constant matrix have negligible values at its boundaries . in addition , the calculation of the forces must be converged with respect to electronic @xmath70-point sampling , and to enhance this convergence we employ fermi smearing . this smearing itself incurs errors , which need to be made negligible . in summary , the phonon frequencies must be converged with respect to atom displacement , supercell size , @xmath70-point sampling and fermi smearing . rather than insisting that every single frequency be converged , it is more convenient to require that the geometric mean frequency @xmath93 be converged . this quantity is defined by the equation : @xmath94 where @xmath95 is the phonon frequency of branch @xmath23 at wave vector @xmath96 , and @xmath97 is the number of branches times total number of * q * points in the sum . it is useful to work with @xmath93 , because it is directly related to the harmonic free energy of lattice vibrations , which , well above the debye temperature , is equal to @xmath98 per atom . our aim is to have @xmath93 converged to better than @xmath99 % with respect to all technical parameters . tests on small supercells show that an atomic displacement of @xmath100 is small enough to ensure that anharmonic errors are well below this tolerance . to test convergence with respect to @xmath70-point sampling and fermi smearing width @xmath101 , we have done extensive tests on a @xmath71 supercell containing 8 atoms ( see table i ) . these tests show that for @xmath102 ev , it is easy to achieve excellent @xmath70-point convergence . but repetition of this at @xmath103 ev reveals that this reduction of @xmath101 causes @xmath93 to change by @xmath104 % . however , further reduction of @xmath101 to @xmath105 ev changes @xmath93 by less than @xmath106 % , so that @xmath93 appears to be adequately converged with respect to @xmath70-points and @xmath101 with a @xmath107 @xmath70-point grid and @xmath103 ev . we then seek convergence with respect to supercell size using this value of @xmath101 and a @xmath70-point density that is reduced in inverse proportion to supercell size . the results indicate that convergence to our required tolerance is achieved with the @xmath108 supercell of 64 atoms . [ fig : vaspphonon ] shows a comparison with experimental data of our calculated phonon frequencies obtained with the @xmath109 supercell of @xmath110 atoms , a @xmath108 @xmath70-point grid , and @xmath103 ev . according to our convergence tests , any discrepancy with experimental frequencies of over 1 % represents a genuine disagreement . the agreement is actually very satisfactory over most of the brillouin zone , with typical discrepancies being @xmath111 % . however , there is a region around the h point @xmath112 , where there are discrepancies of @xmath113 % . this same h - point problem has been noted by previous authors who dealt with transition metals ( mo and nb ) and used pseudopotential - based methods @xcite . the origin of this sharp dip in the phonon dispersion curve of mo at point h has been related to the nesting of electronic states near the fermi level @xcite ( see figures in sec . [ sec : electronic_states ] ) , so it is likely that by reducing the fermi smearing and increasing the number of @xmath70-points the agreement with experiments would be improved . in any case , we believe that since the discrepancies are rather localized in @xmath70-space , they will have only a weak effect on the thermodynamic properties of the system . in summary , our tests show that : ( i ) neither lda nor gga perfectly reproduces the experimental @xmath14 k pressure - volume curve , but the volume given by gga deviates by only a small and almost constant amount of @xmath114 % from the experimental value over the pressure range @xmath73 gpa ; ( ii ) with gga , the inclusion of 4s states in the valence set makes a negligible difference to the @xmath88 curve ; ( iii ) comparisons of paw and fp - lapw confirm the accuracy of paw for both @xmath88 and the electronic dos , and in particular paw accurately reproduces the well - known pressure induced shift of s - p bands relative to d bands ; ( iv ) gga gives rather accurate phonon frequencies over most of the brillouin zone . this evidence provides a firm basis for our calculations on the high - pressure melting of mo , which employ the paw technique with gga(pbe ) exchange - correlation , and with 4p states but not 4s states in the valence set . we begin this section by presenting our _ ab initio _ calculations of the melting curve of mo using the reference coexistence technique ( see sec . [ sec : reference_coexistence ] ) ; our results for the volume and entropy changes on melting as a function of pressure are also reported . in sec . [ sec : structure_changes ] , we outline our calculations of the atomic and electronic structures of the b.c.c . solid and the liquid . we start by determining the _ ab initio _ @xmath4 at a pressure close to zero . at this pressure , we have experimental values for @xmath4 ( 2883 k ) @xcite and the volumes per atom of coexisting solid and liquid ( 16.34 and 17.04 / atom ) @xcite . we use this information to set the temperature and volume / atom of the _ ab initio _ m.d . simulations that we performed to fit the parameters of the reference model . these m.d . simulations employed systems of 125 atoms with @xmath115-point sampling , and were done at @xmath116 k , @xmath117 @xmath118/atom ( solid ) and @xmath119 k , @xmath117 @xmath118/atom ( liquid ) . the starting configuration of the solid was produced by equilibrating the system , which initially was in the perfect crystal configuration , to temperature @xmath120 k. for the liquid , we produced the starting configuration by raising the temperature of the perfect crystal to @xmath121 k ( more than twice the experimental @xmath4 ) and then rethermalizing it again to @xmath122 k. we checked that the system was in the liquid state by monitoring the time - dependent mean - squared displacement . the durations of the simulations were about @xmath123 ps for the solid and @xmath124 ps for the liquid . a set of 100 solid and liquid configurations from these _ ab initio _ simulations was then used to fit the reference model by varying the eam parameters to minimize the dimensionless quantity @xmath63 and the pressure difference @xmath50 ( see sec . [ sec : reference_coexistence ] ) . the resulting eam parameters are reported in table ii . the very small value @xmath125 indicates a good quality of fit , and the @xmath50 values of 0.3 gpa and 0.7 gpa for the liquid and solid , respectively , were also satisfactory ( these @xmath50 give contributions of only @xmath126 and @xmath127 ev / atom to @xmath128 ) . we then performed reference coexistence simulations , and found that the solid and liquid remain in stable coexistence over periods of @xmath129 ps at @xmath130 gpa and @xmath131 k. to illustrate this , we show in fig . [ fig : slice ] the density profile obtained by calculating the number of atoms in slices parallel to the solid - liquid interface . the solid is immediately recognisable from the regular oscillations with a repeat distance of 2.75 ( equal to @xmath132 times the lattice parameter of the bcc lattice , corresponding to the distance between nearest neighbours ) , whereas the density profile is flat in the liquid region . finally , we corrected for the difference between _ ab initio _ and reference energy functions , obtaining a final _ ab initio _ k at @xmath130 gpa . the values of @xmath134 ( where @xmath135 ) and of @xmath136 of the liquid and solid used to make these corrections are reported in table iii . we comment below on the volume and entropy of melting . we now use the reference model to obtain a first estimate of the melting curve at higher pressures . reference coexistence simulations performed with the eam parameters from the fit at @xmath137 gpa showed that at @xmath138 gpa , the reference @xmath4 is 5392 k , the volumes of coexisting solid and liquid being 13.30 and 13.53 @xmath118/atom . correcting for the differences between _ ab initio _ and reference energy functions , we obtain the _ ab initio _ melting temperature @xmath139 k at @xmath138 gpa . we note that the correction to @xmath4 is considerably greater than at @xmath140 gpa , but we believe it is still small enough for the first - order correction scheme to remain valid , and this is confirmed by subsequent results ( see below ) . however , when we repeated this procedure at @xmath141 gpa , still using the reference model fitted at @xmath137 , we found that the corrections need to go from reference to _ ab initio _ @xmath4 were even larger , and we considered it essential to refit the reference model . rather than attempting to do this refit at @xmath141 gpa , we returned to @xmath142 gpa , where our knowledge of the _ ab initio _ @xmath4 is reasonably secure . the refitting at @xmath142 gpa produced the new parameters reported in table ii ( @xmath143 gpa ) . this new reference model , when used in coexistence simulations at @xmath144 gpa , yielded the reference @xmath145 k , and a corrected _ ab initio _ value @xmath146 k. we regarded the size of this correction as acceptably small . in a similar way , when we performed calculations at @xmath147 gpa , using the reference model fitted at 90 gpa , the corrections were unacceptably large , and we performed a refit at 160 gpa . this refitted reference model required only rather small corrections when used at 270 and 380 gpa . the reference and _ ab initio _ @xmath4 as a function of pressure from this full set of calculations are reported in fig . [ fig : melting ] . we find that the _ ab initio _ @xmath4 values can be very well fitted with the so - called simon equation @xcite @xmath148 , with @xmath149 k , @xmath150 gpa and @xmath151 . the resulting @xmath152 melting temperature of 2894 k is very close to the experimental value of 2883 k. using the simon equation , we can obtain the melting slope @xmath2 at any pressure . at @xmath152 , we find @xmath153 k gpa@xmath154 , which agrees closely with the experimental value @xcite of @xmath155 k gpa@xmath154 . also shown in fig . [ fig : melting ] is the point on the melting curve at @xmath156 gpa estimated from the shock data of hixson _ et al . _ @xcite , which is close to our melting curve . the diamond anvil cell ( dac ) measurements of errandonea _ et al . _ @xcite differ greatly from our results , since their @xmath2 is essentially zero over most the range from 0 to 100 gpa . previous theoretical melting curves @xcite for mo , also shown in the figure , are in general agreement with our results , though there are substantial quantitative differences . the comparison of all these experimental and theoretical results raises important issues , which will be discussed in sec . [ sec : discussion ] . the entropy and volume of fusion of the reference model , denoted by @xmath40 and @xmath157 , are straightforward to calculate . from the reference coexistence simulations , we have @xmath158 pairs lying on the reference melting curve . we then perform separate single - phase m.d . simulations of the solid and liquid reference systems at chosen temperatures , using systems of 3375 atoms , adjusting the volumes in each case to give the appropriate @xmath5 . the difference of the resulting volumes give us @xmath157 . at the same time , the simulations give the internal energy @xmath159 , from which we obtain the enthalpies @xmath160 of the two phases . then the difference of enthalpies @xmath68 gives us the entropy difference , since @xmath161 . we find that @xmath40 is almost constant along the reference melting curve , going from @xmath162 @xmath163/atom at @xmath164 to @xmath165 @xmath163/atom at @xmath166 gpa . since the reference system mimics the _ ab initio _ system rather closely , we assume that @xmath167 for the _ ab initio _ system is essentially the same as @xmath40 . to obtain the _ ab initio _ volume of fusion @xmath168 , we use the clausius - clapeyron relation @xmath169 . our @xmath7 results for the reference and _ ab initio _ systems as a function of @xmath5 are reported in fig . [ fig : changevol ] . we note that in both cases the fractional volume change @xmath170 decreases smoothly from @xmath114 % at @xmath152 to @xmath171 % at 400 gpa . we noted in the introduction that theories of the melting of b.c.c . transition metals sometimes assume @xcite that melting is associated with a significant change of coordination number , so that the electronic density of states should also change markedly . in order to test these ideas for mo , we have calculated the radial distribution function @xmath172 of mo from a series of aimd simulations of the solid and the liquid performed at @xmath173 gpa . the simulations were all done with 125 atoms , and had a typical duration of 2 ps after equilibration . [ fig : gr ] reports our calculated @xmath172 at @xmath174 , 4000 and 6000 k ( solid state ) , and at @xmath175 k ( liquid state ) ; we recall ( fig . [ fig : melting ] ) that our calculated @xmath4 at this pressure is @xmath176 k. at 2000 k , the shells containing 8 first neighbours and 6 second neighbours at distances of @xmath177 and 2.81 are clearly separated , and @xmath172 goes to zero at @xmath178 , between the second and third shells . however , at @xmath179 k , the first and second shells have already merged , though the shoulder due to the second shell is clearly visible . the @xmath172 of the solid at @xmath180 k presents similar trends to those shown at 4000 k , the main difference being that the valley between the second and third atomic shells now clearly moves upwards from zero . the change in @xmath172 in going from the solid at @xmath180 k to the liquid at @xmath181 k is substantial for @xmath182 , with the well defined peaks due to third and higher neighbours in the solid becoming heavily broadened . however , the peak closest to the origin does not suffer a large change . we define the coordination number @xmath183 in the usual way as @xmath184 , with @xmath185 the bulk number density and @xmath186 the distance of the first minimum . the @xmath186 values at the four temperatures are 3.41 , 3.41 , 3.41 and 3.31 , and the resulting @xmath183 values are 14 , 14 , 14 and 13.35 . apart from the expected increase in disorder , the main change on going from solid to liquid is thus a slight _ decrease _ in coordination number . we comment further on this in sec . [ sec : discussion ] . turning now to the electronic density of states ( dos ) , we present first our aimd results for @xmath5 in the range @xmath187 gpa at a series of temperatures , the simulations being performed on a system of 125 atoms with @xmath115-point sampling . the typical duration of these simulations was 2 ps after equilibration and the dos were calculated by averaging over 150 different configurations . we report in fig . [ fig : dosmelt ] the calculated dos at the thermodynamic states given by @xmath188 , @xmath189 , @xmath190 for the solid , and @xmath191 for the liquid ( units of gpa and k ) . ( our _ ab initio _ @xmath4 for @xmath192 gpa are in the range @xmath193 k. ) we have checked in each case that the system is in the solid or liquid states by looking at the mean - squared displacement and structure factor . we note the progressive broadening of the dos peaks with increasing thermal disorder in the solid , an effect which continues further in the liquid . the fermi - level value of the dos increases slightly on melting . as far as occupied states are concerned , melting appears to cause a slight redistribution of d - states from lower in the band to the region of the fermi level . in the right - hand panel of fig . [ fig : dosmelt ] , we compare our aimd results for the electronic dos at the solid state - point @xmath194 and the liquid state - point @xmath195 ( units of gpa and k ) , which are just below and just above our calculated melting curve . we note the rather minor changes caused by melting . interestingly , the fermi - level value of the dos is almost identical in the two phases at this pressure . the relationship of these results with earlier work on the electronic structure of liquid transition metals will be discussed in the following section . at the start of this paper , we emphasized the large discrepancies between melting curves of transition metals derived from static compression and shock measurements , and we mentioned that previous dft work on mo supports the shock measurements . the present work fully confirms that the melting curve predicted by dft in the pbe approximation for exchange - correlation energy lies far above the static compression measurements , but at high pressures is consistent with the shock data . this confirmation is important , because of deficiencies or uncertainties in previous dft work . the reliability of the present calculations is supported by our close agreement with the experimental @xmath152 values of both the melting temperature @xmath4 and the melting slope @xmath2 . our melting curve is below the theoretical curve of moriarty @xcite by @xmath196 k at @xmath152 , and this difference increases with increasing @xmath5 . however , this is not surprising , since the generalized pseudopotential model that he used is known to disagree with experimental phonon frequencies , and because he included thermal electronic excitations only approximately . in the present work , we have taken pains to verify the accuracy of the phonon frequencies given by our methods , and thermal electronic excitations are fully included within our dft framework . perhaps more surprising is that our melting curve agrees closely with that obtained by belonoshko _ _ @xcite using direct dft m.d . simulation . this is unexpected , since they believed that their melting curve suffered from a substantial superheating error of @xmath12 % . the close agreement suggests that they may have been unduly pessimistic , and this point deserves further investigation . we included in fig . [ fig : melting ] the results of belonoshko _ et al . _ @xcite and of verma _ et al . _ @xcite obtained by the dislocation - mediated theory of melting @xcite . it is not clear to us whether one can expect a theory of melting based exclusively on the properties of the solid to be fully reliable . one of the problems with this approach is that the predicted melting curves rely on thermodynamic data that may not be reliably known . the rather large differences between the two melting curves based on the dislocation theory may be indicative of the limited reliability of this approach . the change of volume on melting of @xmath104 % given by our calculations is small , but still much greater than the volume change implied by the static compression values @xcite of @xmath2 . arguments in favour of a very low volume change based on a significant increase of coordination number on going from b.c.c . solid to melt appear to be incorrect , according to our dft m.d . calculations of the radial distribution function @xmath172 . we find only rather minor differences between @xmath172 for high-@xmath1 solid and melt . in particular , there is actually a slight _ decrease _ in coordination number from 14 to @xmath197 on melting , so that the liquid is slightly less close packed than the solid . we comment that ideas based on hard - sphere packing are likely to be misleading , since the repulsive interactions between mo atoms at high @xmath1 are rather soft ( see below ) . we mentioned in the introduction the recent theory of ross _ et al . _ @xcite , according to which a very low melting slope is expected for b.c.c . transition metals . the theory invokes the well known transfer of electrons from s - p to d states with increasing compression , and the fact that this transfer depends on crystal structure . in applying this theory to the melting of mo , the authors estimated the effective number of d electrons @xmath198 by treating the high - temperature solid as a perfect b.c.c . crystal and the liquid as a perfect f.c.c . crystal . they also assumed that a change of @xmath198 on melting will be associated with a change of d - band width . they found that the changes of electronic structure stabilize the liquid relative to the solid , and yield a major reduction of @xmath4 . in considering this theory in the light of the results we have presented , it is important to appreciate that our calculations are all based on an accurate implementation of dft . as in all simulations using born - oppenheimer dft m.d . , the vasp code recalculates the entire self - consistent electronic structure at every time step of the time evolution . as described in sec . [ sec : electronic_states ] , we have gone to considerable lengths to show that the paw implementation of dft used in our m.d . yields results for the electronic dos which are almost indistinguishable from those given by the fp - lapw technique , which is one of the most accurate available . this means that all the effects that enter the theory of ross _ et al . _ @xcite , including s - p to d electron transfer and changes of d - band width , are fully included in our simulations . nevertheless , we do not obtain the very low melting slope that they predict . the reason for this is presumably that their treatment of the high-@xmath1 solid and the liquid as perfect crystals is incorrect . as we have seen , their assumption of a large structural change on melting also appears to be questionable . this point is reinforced by our finding that the electronic dos changes only slightly on melting , especially at high @xmath5 . we end this discussion by commenting on the embedded - atom model ( eam ) used as a reference system in determining the melting curve . it is an important finding of this work that the eam is able to mimic accurately the dft total - energy function of solid and liquid close to coexistence . this does not mean , however , that we accept the melting curve of the reference model as the true melting curve . this could be dangerous , because we might then miss d - band electronic effects that were not explicitly included in the model . however , in our procedure , any such effects are automatically picked up in the corrections that we apply , since these explicitly account for free energy differences between the reference and dft systems . we note in passing that the fitting of our reference model yields parameters that resemble those we found in our earlier work on the melting of fe @xcite . in particular , the inverse - power repulsive potential in our present eam model has an exponent @xmath30 close to 6 at low @xmath5 , decreasing to @xmath113 at high @xmath5 . for comparison , the fitted eam in our fe work had @xmath199 , which is very similar . we are currently investigating the systematic behaviour of eam parameters in solid and liquid transition metals at high @xmath5 and @xmath1 , and we hope to report on this elsewhere . the main conclusions from this work are as follows . our dft calculations of the melting curve of mo up to 400 gpa fully confirm earlier work showing that the dft melting curve is consistent with shock data , but is far above the melting curve given by static compression experiments . our calculations indicate that at high @xmath5 there are only minor changes of both atomic and electronic structure on going from the high - temperature b.c.c . solid to the melt . suggested mechanisms for an anomalously low melting slope @xmath2 of mo based on transfer of electrons from s - p states to d states appear to be incompatible with the present dft calculations . this tends to confirm earlier suggestions @xcite that the transition identified as melting in high-@xmath5 static compression experiments may not be true thermodynamic melting . the work was supported by epsrc grant ep / c534360 , which is 50% funded by dstl(mod ) , and by nerc grant ne / c51889x/1 . the work was conducted as part of a euryi scheme award to da as provided by epsrc ( see www.esf.org/euryi ) . in order to ensure energy convergence to within 1 mev / atom in the wien2k calculations , we use the following settings : tolerance for energy convergence in the self - consistency cycle : @xmath202 ev ; muffin - tin radius ( @xmath203 ) ranges from @xmath204 for the largest atomic volume to @xmath205 for the smallest ; plane - wave cut - off @xmath206 ; maximum angular momentum component for partial waves used inside atomic spheres @xmath207 ; number of wavevectors in full brillouin zone for @xmath70-point sampling @xmath208 ; @xmath209 and @xmath82 states are treated as valence states throughout . [ fig : vasp - lda - pbe ] : comparison of lda and gga pressure @xmath5 as function of volume @xmath52 for b.c.c . mo from the paw method with different exchange - correlation functionals and valence sets . long - dashed and solid lines ( practically coincident ) show gga results with and without @xmath77 states in the valence set . short - dashed and dotted lines show lda results with and without @xmath77 states in the valence set . dots show experimental results @xcite . [ fig : wien2kvasp ] : comparison between paw and fp - lapw results for the gga(pbe ) and lda(ca ) approximations for @xmath13 . solid and dashed curves show gga(pbe ) and lda(ca ) fp - lapw results , respectively ; short - dashed and dotted curves show gga(pbe ) and lda(ca ) paw calculations , respectively . solid dots show experimental data @xcite . [ fig : doswien2kvasp ] : density of electronic states obtained with the paw ( dashed line ) and fp - lapw ( solid line ) at zero temperature and 0 , 150 and 300 gpa . fermi energies are shifted to zero ( dotted line ) . [ fig : bandstructure ] : fp - lapw calculation of the energy bands of mo at 0 and 300 gpa ( left and right panels respectively ) . the @xmath212 valence band ( energy between -8 and -7 ev at the @xmath115 point ) rises in energy more quickly than @xmath25 valence bands with increasing pressure . + fig . [ fig : vaspphonon ] : comparison of calculated ( curves ) and experimental ( solid squares ) phonon dispersion relations of mo at zero pressure . experimental data are from ref . [ fig : slice ] : density profile in simulation of coexisting solid and liquid mo at @xmath130 gpa , @xmath131 k after 60 ps . the simulation is performed with the embedded - atom reference model on a system of 6750 atoms . [ fig : melting ] : calculated _ ab initio _ melting curve ( filled circles and solid line ) of this work compared with previous results : generalized pseudopotential calculations of moriarty @xcite(dotted line ) , dislocation - mediated models of belonoshko _ et al . _ @xcite(long - dashed line ) and verma _ et al . _ @xcite(dashed - dotted line ) ; experimental shock - wave @xcite and dac @xcite measurements are shown with empty squares and triangles , respectively . filled and inverted - empty triangles show solid and liquid _ ab initio _ molecular dynamics calculations of belonoshko _ et al . _ @xcite , respectively . empty circles show results of this work obtained with the eam model without free - energy corrections . [ fig : changevol ] : _ ab initio _ fractional volume change on melting of mo as a function of pressure . solid and dashed curves : present work , with and without free - energy correction , respectively . [ fig : gr ] : calculated radial distribution function of mo for : solid at @xmath213 gpa and @xmath174 k ( solid line ) , solid at @xmath214 gpa and @xmath179 k ( dotted line ) , solid at @xmath215 gpa and @xmath180 k ( long - dashed line ) and liquid at @xmath216 gpa and @xmath175 k ( short - dashed line ) . + fig . [ fig : dosmelt ] : density of valence electronic states of mo at finite temperature and on melting . left : solid at @xmath217 gpa and @xmath14 k ( dotted line ) , solid at @xmath218 gpa and @xmath174 k ( short - dashed line ) , solid at @xmath219 gpa and @xmath220 k ( solid line ) and liquid at @xmath221 gpa and @xmath222 k ( long - dashed line ) . right : solid at @xmath223 gpa and @xmath224 k ( solid line ) and liquid at @xmath92 gpa and @xmath225 k ( dashed line ) . fermi energy levels are shifted to zero .

we report _ ab initio _ calculations of the melting curve of molybdenum for the pressure range @xmath0 gpa . the calculations employ density functional theory ( dft ) with the perdew - burke - ernzerhof exchange - correlation functional in the projector augmented wave ( paw ) implementation . we present tests showing that these techniques accurately reproduce experimental data on low - temperature b.c.c . mo , and that paw agrees closely with results from the full - potential linearized augmented plane - wave implementation . the work attempts to overcome the uncertainties inherent in earlier dft calculations of the melting curve of mo , by using the `` reference coexistence '' technique to determine the melting curve . in this technique , an empirical reference model ( here , the embedded - atom model ) is accurately fitted to dft molecular dynamics data on the liquid and the high - temperature solid , the melting curve of the reference model is determined by simulations of coexisting solid and liquid , and the _ ab initio _ melting curve is obtained by applying free - energy corrections . our calculated melting curve agrees well with experiment at ambient pressure and is consistent with shock data at high pressure , but does not agree with the high pressure melting curve deduced from static compression experiments . calculated results for the radial distribution function show that the short - range atomic order of the liquid is very similar to that of the high-@xmath1 solid , with a slight _ decrease _ of coordination number on passing from solid to liquid . the electronic densities of states in the two phases show only small differences . the results do not support a recent theory according to which very low @xmath2 values are expected for b.c.c . transition metals because of electron redistribution between s - p and d states .

here we present the basic equations of the problem . in the space - time of the closed cosmological friedmann models with space - time metric @xmath0,\ ] ] where @xmath1 , the action is assumed to be @xmath2\sqrt{-g}\,d\omega,\ ] ] here @xmath3 is a scalar curvature , @xmath4 is a potential of the scalar quantum field with a mass @xmath5 , ( @xmath6 ) , @xmath7 is a cosmological @xmath8 - term . the equation for @xmath9 has the form @xmath10 . below let us set the scalar field as radial standing waves with time - dependent amplitudes @xmath11 note that the functions @xmath12 do not have singularities on the poles @xmath13 for integer numbers @xmath14 , and that they are orthonormal on the coordinate @xmath15 in the measure defined by metric ( [ doc1 ] ) , i.e. , @xmath16 now the action ( [ doc2 ] ) can be integrated over the angular variables , and we rewrite it after replacement of @xmath3 by its value through @xmath17 as @xmath18dt,\ ] ] where @xmath19 these redefinitions are made to get @xmath20 and @xmath21 only in front of the integral , and the integrand to be free of both constants . in this formula , the point denotes the derivative with respect to @xmath22 , and one supposes that all @xmath23 , i.e. , we assume that waves with wavelength comparable to the diameter of the universe are negligible for the dynamics . the waves packet @xmath24 ensures the homogeneity of the model during expansion with a high accuracy . after introduction of generalized momenta the hamiltonian function can be presented as @xmath25 here @xmath26 are a generalized momentum and a coordinate for the universe as entire ; @xmath27 and @xmath28 are the same for each radial mode of the field . corresponding to eq . ( [ doc6 ] ) the wheeler - dewitt equation for the wave function of the universe ( wf ) has the form @xmath29 where the superpotential @xmath30 is given as @xmath31 .\ ] ] here we are neglecting @xmath32 in comparison with the sum residual in eq . ( [ doc8 ] ) since @xmath23 . this is justified because in the short - wavelength approximation we are applying here , the mass of the particles does not have much influence . the factor ordering parameter @xmath33 @xcite is also introduced in eq . ( [ doc8 ] ) . we are going to search a solution of equation ( [ doc7 ] ) in semiclassical approximation @xmath34 where @xmath35 is a classical action . in this approximation we have @xmath36 this nonlinear equation of the first order is similar to the hamilton - jacobi equation of analytical mechanics . for its solution one can take advantage of the complete integral @xcite . it can be found from the system of characteristic equations @xcite . as the eq . ( [ doc10 ] ) supposes the solution for @xmath37 , the scale factor becomes an argument along the characteristics . from here we have @xmath38 here @xmath30 is indicated by ( [ doc8 ] ) , and @xmath39 now the evolution of the wf @xmath40 goes along a characteristic . turning points are the roots of @xmath41 . they separate classically available and forbidden regions along the characteristic eq . ( [ doc11 ] ) @xcite . a similar approach for the strongly anisotropic wf in a bianchi type i model is presented in ref . @xcite , and for a closed friedmann model in ref . @xcite . in the region under the barrier the semiclassical solution can be found by an ansatz of the form @xmath42 the equation for @xmath43 can be obtained from eq . ( [ doc10 ] ) by changing @xmath44 . thus , as distinct from eq . ( [ doc11 ] ) the system of the characteristics has the form @xmath45 @xmath46 let us investigate how amplitudes of short wavelength change at the above mentioned evolution of the wf . equations for @xmath27 and @xmath47 follow from ( [ doc11 ] ) @xmath48 where @xmath49 this system is an essentially non - linear one since co - factors of @xmath27 and @xmath47 in the right hand sides of the equations depend on @xmath17 , too . however , the presence of the large parameter @xmath50 allows finding a solution . following from the system ( [ doc18 ] ) , the equation for the generalized momenta @xmath27 is @xmath51 where @xmath52 . here @xmath53 is an unknown function according to eq . ( [ doc18 ] ) . it is easy to show that the following solution @xmath54 satisfies the last equation with the high accuracy of @xmath55 . and correspondingly , @xmath56 where @xmath57 is an arbitrary constant , and the phase of the solution is chosen from the requirement of amplitude finiteness at @xmath58 . the mentioned accuracy of solutions requires the assumption that the logarithmic derivative @xmath59 does not give a large parameter @xmath14 . the last statement follows from the fact that each mode of wave enters into eq . ( [ doc18 ] ) for @xmath60 as a term @xmath61 it is clear from substitution eqs . ( [ doc20 ] ) , ( [ doc21 ] ) in ( [ doc22 ] ) . with taking into account ( [ doc22 ] ) after rewriting of the amplitude @xmath62 , we have for @xmath60 the following expression @xmath63},\ ] ] further , we suppose that @xmath64 and we have in the superpotential @xmath65.\ ] ] values of @xmath30 and @xmath60 define an action of wf @xmath66 according to the system of characteristics ( [ doc11 ] ) . because of the large number of short waves , the sum in eq . ( [ doc22b ] ) can be taken as fast oscillations average with an adequate accuracy . then the expression for the superpotential gets the form @xmath67,\ ] ] the existence of an inner classically available region requires a realization of the condition @xmath68 in the mentioned interval @xmath69 . ( let us remember see eq . ( [ doc6 ] ) that the total energy @xmath70 vanishes for the closed universe . ) the potential barrier should begin at @xmath71 , i. e. @xmath72 at @xmath73 . let us notice that the vanishing of the factor ordering parameter ( @xmath74 ) allows us to get an inner classically available region only at the expense of the sum @xmath75 . but in that case , a potential barrier necessary for a further development is absent . so , we choose the factor ordering @xmath76 in the following . owing to the amplitude of short waves begin to increase quickly under the barrier , we assume that their total energy is negligible in the considered region @xmath68 , and the evolution of the universe is determined only by the first two terms in the r.h.s . ( [ doc22c ] ) . in this case the dynamics of the amplitudes of the short waves can be given explicitly . the phase in eq . ( [ doc20 ] ) is @xmath77 here , @xmath78 can be obtained from the condition @xmath79 = 0,\qquad a_0=\frac{1}{2}.\ ] ] a case when the total energy of the modes is not negligible in the expressions for @xmath30 and @xmath60 can be similarly considered . but then the final expression becomes essentially complicated . the superpotential @xmath30 is positive at @xmath80 . this means that the evolution of the wf occurs under the barrier . at this range , the problem is defined by the system of characteristics ( [ doc13 ] ) , and we have for @xmath27 the following equation @xmath81 this equation has an inverse sign at first term on the right in comparison with eq . ( [ doc19 ] ) for the classically allowed region . it changes the solution in principle , but for a large parameter @xmath50 we can find a solution by a method similar to the one used for eq . ( [ doc19 ] ) @xmath82 @xmath83 @xmath84 here the conditions of continuity of @xmath27 and @xmath47 at the point @xmath71 were used for the selection of the constants of integration . both in section 3 and here , the correctness of the solution ( [ doc22e ] ) is determined by the requirement of @xmath85 . this requirement is checked by substituting of the obtained solution in @xmath86 of eq . ( [ doc13 ] ) . there @xmath87 in the last formula the term containing the factor ordering , that plays a part only close to zero , is ignored . because of the smallness of the sum @xmath88 under the barrier this term can be neglected in the expression for @xmath86 , and we suppose with a necessary accuracy @xmath89 at that @xmath17 the phase @xmath90 has the form @xmath91 and leaving only growing modes in solutions at @xmath80 we obtain @xmath92 in that solution the amplitude of short waves catastrophically increases with growth of @xmath14 but not under the exponential law . the latter is realized in the case of a slow variation of @xmath86 in eq . ( [ doc22e ] ) . it is similar to the rubakov - effect of catastrophic particle creation , see ref . @xcite . a final value of an amplification of the amplitude of short waves is determined by a time " of the wf stay under the barrier . let us notice that we consider a self - consistent problem of amplification of short waves ( an analogue of particle creation ) . it means that the energy of the increasing modes under the barrier controls the form of the barrier itself ( see fig . ( 50,160 ) ( -200,-280 ) let us consider the equation for the action of the wf under the barrier ( [ doc13 ] ) @xmath93 only a potential " energy of increasing modes is present in the superpotential @xmath30 , but here again it is not cancelled by their kinetic " energy as it was realized in ( [ doc * ] ) . as it follows from @xmath94 , an external boundary of the tunneling region is determined by the equation @xmath95 the final value of amplification of the amplitude of the modes is defined by substitution of @xmath96 in eq . ( [ doc * * ] ) . further , the euclidean action @xmath97 has a form @xmath98 subject to @xmath99 we have @xmath100 . hence , a probability of tunneling of the wf through the barrier is given as @xmath101 at @xmath102 . the formula @xmath103 is more exact in the case @xmath104 . the self - consistent problem of the evolution of the closed friedmann universe with an amplification of wave perturbations of a scalar field under a barrier is considered . within the framework of a semiclassical approximation of the wheeler - dewitt equation it is shown that the increase of the amplitude of short waves under the barrier controls the barrier shape , i.e. leads to both a reduction of the barrier and fast output of wf from under the barrier . in this sense the problem is self - consistent . the process of tunneling and the increase of the amplitudes of short waves is defined here by the assignment of their initial spectrum in the inner classically allowed region . if it is required after tunneling in the universe to receive a spectrum close to the flat one , the initial spectrum should be almost exponentially suppressed in the area @xmath105 . it is qualitatively equivalent to the presence of an effective temperature in the inner classically allowed region . the authors thank a. a. starobinsky and m. v. sazhin for information about papers of v. a. rubakov et al . and h. kleinert for useful comments . .- j.s . acknowledges financial support from the hsp iii - program . the work is done within the project kr-154 of the international science and technology centre ( istc ) . a. d. linde , `` elementary particle physics and inflationary cosmology '' , nauka , main editorial board for physical and mathematical literature , moscow , 1990 . landau , e.m . livshits , mechanics " . nauka , main editorial board for physical and mathematical literature , moscow , 1988 . e. kamke , the reference book on differential partial equations of the first order " . nauka , moscow , 1966 . gurovich , u. m. imanaliev and i. v.tokareva , _ jetp lett _ * 64 * , no.5 , 329 ( 1996 ) . m. bachmann and h .- j . schmidt , gr - qc/9912068 , phys . d in press . v. folomeev and v. gurovich , gen . * 32 * ( 2000 ) in press . rubakov , _ pisma zh.eksp . teor * 39 * , no . 2 , 89 - 92 ( 1984 ) . lavrelashvili , v.a . rubakov and p.g . tenyakov , _ nuclear physics _ * b299 * , no . 2 , 757 - 796 ( 1988 ) .

the evolution of the closed friedmann universe with a packet of short scalar waves is considered with the help of the wheeler dewitt equation . the packet ensures conservation of homogeneity and isotropy of the metric on average . it is shown that during tunneling the amplitudes of short waves of a scalar field can increase catastrophically promptly if their influence to the metric do not take into account . this effect is similar to the rubakov - effect of catastrophic particle creation calculated already in 1984 . in our approach to the problem it is possible to consider self - consistent dynamics of the expansion of the universe and amplification of short waves . it results in a decrease of the barrier and interruption of amplification of waves , and we get an exit of the wave function from the quantum to the classically available region . = -10.0 mm = 0.0 mm = 0.0 mm

answer set programming ( asp ) , namely logic programming under the answer set semantics @xcite , is a constraint programming paradigm , which has been successfully deployed in many applications @xcite . recently , asp was extended to include constraints to facilitate reasoning with sets of atoms . these constraints include weight constraints @xcite , aggregates @xcite and abstract constraints @xcite . among them , weight constraints and aggregates are the most widely used constraints in practice . in this paper , logic programs with weight constraints and aggregates will be referred to as _ weight constraint _ and _ aggregate programs _ , respectively . the semantics of weight constraint programs , called the _ stable model semantics _ , is well established and implemented in a number of asp solvers @xcite . especially , the results of the asp solver competitions @xcite show that clasp is an efficient solver that implements this semantics . for aggregate programs , various semantics have been proposed @xcite . the one proposed in @xcite ( previously in @xcite ) , called the _ ultimate stable semantics _ , is based on an iterative construction on partial interpretations . the same semantics is reformulated by @xcite and extended to logic programs with arbitrary abstract constraint atoms , which embodies a key concept called _ conditional satisfaction_. since this reformulation is conceptually simpler , as it does not resort to 3-valued logic , in this paper we call this semantics _ conditional satisfaction - based_. among the semantics for aggregate programs , this semantics is known to be the most conservative , in the sense that any answer set under this semantics is an answer set under others , but the reverse may not hold . the relationships of these semantics have been studied in @xcite . in this paper , we refer to the semantics based on conditional satisfaction as the _ answer set semantics_. despite the fact that weight constraint and aggregate programs are among the most popular classes of programs in practice , the relationship among them has not been fully studied , both in semantics and in representation . in this paper , we study the relationship between the stable model semantics and the answer set semantics . we show that for a broad class of weight constraint programs , called _ strongly satisfiable programs _ , the stable model semantics agrees with the answer set semantics . for example , weight constraint programs where weight constraints are upper bound free are all strongly satisfiable . this result is useful in that we are now sure that the known properties of the answer sets also hold for these programs . one important property is that any answer set is a _ well - supported model _ @xcite , ensuring that any conclusion must be supported by a non - circular justification in the sense of @xcite . our study further reveals that for weight constraint programs where the stable model and answer set semantics disagree , stable models may be circularly justified . we then show that the gap between the two can be closed by a transformation , which translates an arbitrary weight constraint program to a strongly satisfiable program so that the answer sets of the original program are exactly the stable models of the translated program . we further demonstrate the precise difference between the two semantics using a more general logic programming framework , logic programs with nested expressions . we propose yet another transformation from weight constraint programs to logic programs with nested expressions which preserves the answer set semantics . we compare this transformation to the one given in @xcite , which is faithful to the stable model semantics . interestingly , the difference is small but subtle : given a weight constraint @xmath0 u$ ] , where @xmath1 and @xmath2 are lower and upper bounds , respectively , and @xmath3 $ ] expresses a collection of literals with weights , while in our transformation the satisfaction of the upper bound is interpreted directly as `` less than or equal to @xmath2 '' , in @xcite the interpretation is by negation - as - failure `` not greater than @xmath2 '' . the observation that the gap between the answer set and the stable model semantics can be closed by a transformation leads to an approach for computing answer sets of aggregate programs using the asp solvers that implement the stable models semantics of weight constraint programs . we propose such an approach where aggregate programs are encoded compactly as weight constraint programs and their answer sets are computed using a stable model solver . we conducted a series of experiments to evaluate this approach . the results suggest that representing aggregates by weight constraints is a promising alternative to the explicit handling of aggregates in logic programs . besides efficiency , another advantage is at the system level : an aggregate language can be built on top of a stable models solver with a simple front end that essentially transforms standard aggregates to weight constraints in linear time . this is in contrast with the state - of - the - art in handling aggregates in asp , which typically requires an explicit implementation for each aggregate . the paper is organized as follows . the next section gives preliminary definitions . in section [ relate ] we relate the stable model semantics with the answer set semantics . we first establish a sufficient condition for the two to coincide , and then discuss their differences . in section [ transform ] , we present a transformation to close the gap between the two semantics , followed by section [ aggr - wprogram ] where we show how to represent aggregate programs by weight constraint programs . further in section [ nested - expression ] , to pinpoint the precise difference between the stable model semantics and the answer set semantics for weight constraint programs , by proposing a transformation from weight constraint programs to logic programs with nested expressions which preserves the answer set semantics , and comparing this with that of @xcite . we implemented a prototype system called alparse and in section [ experiments ] we report some experimental results . section [ conclusion ] concludes the paper . a preliminary version of this paper has appeared as @xcite . the main extensions here include : ( i ) section [ nested - expression ] , where we propose a transformation from weight constraint programs to logic programs with nested expressions which preserves the answer set semantics - this transformation shows exactly what makes the answer set semantics differ from the stable model semantics ; ( ii ) section [ experiments ] , where experiments are expanded including the benchmarks for aggregate programs used in the 2007 asp solver competition @xcite ; and ( iii ) the proofs of all the theorems and lemmas . throughout the paper , we assume a fixed propositional language with a countable set of propositional atoms . a _ weight constraint _ is of the form @xmath4 \ , u \label{w - form}\end{aligned}\ ] ] where each @xmath5 , @xmath6 is an atom , and each atom and not - atom ( negated atom ) is associated with a _ weight_. atoms and not - atoms are also called _ literals _ ( the latter may be emphasized as _ negative _ literals).the literal set of a weight constraint @xmath7 , denoted @xmath8 , is the set of literals occurring in @xmath7 . the numbers @xmath1 and @xmath2 are the _ lower _ and _ upper bounds _ , respectively . the weights and bounds are real numbers . either of the bounds may be omitted in which case the missing lower bound is taken to be @xmath9 and the missing upper bound by @xmath10 . a set of atoms @xmath11 satisfies a weight constraint @xmath7 of the form ( [ w - form ] ) , denoted @xmath12 , if ( and only if ) @xmath13 , where @xmath14 @xmath11 satisfies a set of weight constraints @xmath15 if @xmath16 for every @xmath17 . a weight constraint @xmath7 is _ monotone _ if for any two sets @xmath18 and @xmath19 , if @xmath20 and @xmath21 , then @xmath22 ; otherwise , @xmath7 is _ nonmonotone_. there are some special classes of nonmonotone weight constraints . @xmath7 is _ antimonotone _ if for any @xmath18 and @xmath19 , @xmath22 and @xmath21 imply @xmath20 ; @xmath7 is _ convex _ if for any @xmath18 and @xmath19 such that @xmath21 , if @xmath20 and @xmath22 then for any @xmath23 such that @xmath24 we have @xmath25 . a _ weight constraint program _ is a finite set of _ weight rules _ of the form @xmath26 where each @xmath27 is a weight constraint . given a ( weight ) rule @xmath28 of the above form , we will use @xmath29 to denote @xmath30 and @xmath31 the conjunction of the weight constraints in the body of the rule . we will use @xmath32 to denote the set of the atoms appearing in a program @xmath33 . weight constraint programs are often called _ lparse programs _ , which generally refer to the kind of non - ground , function - free logic programs one can write based on the lparse syntax . these programs are grounded before calling an asp solver . in this paper , for the theoretical study we always assume a given weight constraint program is ground . given a weight constraint program @xmath33 , if the head of each rule in @xmath33 is of the form @xmath34~1 $ ] where @xmath35 is an atom , then @xmath33 is said to be _ basic_. if , in addition , all the weight constraints in the bodies of rules in @xmath33 are of the form @xmath36~1 $ ] , where @xmath1 is a literal , then we have a normal program . we will simply write a weight constraint @xmath36~1 $ ] as @xmath1 , since they are equivalent in terms of satisfaction . as pointed out by @xcite , negative weights and negative literals are closely related in that they can replace each other and that one is inessential when the other is available . negative weights can be eliminated by applying the following transformation @xcite : given a weight constraint @xmath7 of the form ( [ w - form ] ) , if @xmath37 , then replace @xmath38 with @xmath39 and increase the lower bound to @xmath40 and the upper bound to @xmath41 ; if @xmath42 , then replace @xmath43 with @xmath44 and increase the lower bound to @xmath45 and the upper bound to @xmath46 . for instance , the weight constraint @xmath47 ~1 $ ] can be transformed to @xmath48 ~4 $ ] note that this transformation is satisfaction - preserving , in the sense that for any weight constraint @xmath7 and set of atoms @xmath11 , @xmath16 iff @xmath49 , where @xmath50 is obtained by applying the transformation . from now on , we assume that negative weights are always eliminated by the above transformation . the stable models of weight constraint programs are defined using the reduct of weight constraints , which is defined as follows : the _ reduct _ of a weight constraint @xmath7 of the form ( [ w - form ] ) w.r.t . a set of atoms @xmath11 , denoted by @xmath51 , is the weight constraint @xmath52\end{aligned}\ ] ] where @xmath53 . let @xmath33 be a weight constraint program and @xmath11 a set of atoms . the reduct of @xmath33 w.r.t . @xmath11 , denoted @xmath54 , is defined by @xmath55 [ smodels - semantics ] let @xmath33 be a weight constraint program and @xmath56 . @xmath11 is an _ stable model _ of @xmath33 iff the following two conditions hold : 1 . @xmath57 , 2 . @xmath11 is the deductive closure of @xmath54 . note that @xmath54 is a monotone basic weight constraint program . the deductive closure of such a program can be constructed using the operator @xmath58 defined in @xcite . let @xmath33 be a monotone basic weight constraint program . the operator @xmath58 is defined as @xmath59 we note that , for a monotone program @xmath33 , the operator @xmath58 is monotone with respect to @xmath19 . then we have the lemma below . [ sm - fixpoint ] given a weight constraint program @xmath33 , a set of atoms @xmath11 is a stable model of @xmath33 iff @xmath57 and @xmath60 . let @xmath33 be a weight constraint program and @xmath11 a set of atoms . @xmath54 is a monotone program . the deductive closure of @xmath54 is the least fixpoint of @xmath61 . then the lemma follows from definition [ smodels - semantics ] . following @xcite , we define the syntax and semantics for aggregate programs below . an aggregate is a constraint on a set of atoms taking the form @xmath62 where @xmath63 is an _ aggregate function_. the standard aggregate functions are those in @xmath64sum , count , avg , max , min@xmath65 . the set @xmath66 is called an _ intensional set _ , where @xmath67 is a predicate , and @xmath68 is a variable which takes value from a set @xmath69 , called the _ variable domain_. the relational operator @xmath70 is from @xmath71 and @xmath72 is either a variable or a numeric constant . the _ domain _ of an aggregate @xmath73 , denoted @xmath74 , is the set of atoms @xmath75 . the size of an aggregate is @xmath76 . for an aggregate @xmath73 , the intensional set @xmath66 , the variable domain @xmath77 , and the domain of an aggregate @xmath74 can also be a multiset which may contain duplicate members . let @xmath11 be a set or multiset of atoms . @xmath11 is a _ model _ of ( satisfies ) an aggregate @xmath73 , denoted @xmath78 , if @xmath79 holds , otherwise @xmath11 is not a model of ( does not satisfy ) @xmath73 , denoted @xmath80 . for instance , consider the aggregate @xmath81 , where @xmath82 . for the sets @xmath83 and @xmath84 , we have @xmath85 and @xmath86 . for the multiset @xmath87 , we have @xmath88 . an aggregate program is a set of rules of the form @xmath89 where @xmath90 is an atom and @xmath91 are aggregates.s could also be atoms or negative atoms . here we focus on aggregates . the results can be extended to the general case , where the atoms and negative atoms are treated exactly the same as that in normal logic programs @xcite . ] for a rule @xmath28 of the form ( [ aggr - rule - form ] ) , we use @xmath29 and @xmath31 to denote @xmath90 and the set @xmath92 , respectively . the definition of _ answer set _ of aggregate programs is based on the notion of _ conditional satisfaction_. [ aggr - cond - sat ] let @xmath73 be an aggregate and @xmath18 and @xmath19 two sets of atoms . @xmath18 conditionally satisfies @xmath73 , w.r.t . @xmath19 , denoted @xmath93 , iff @xmath94 and for every set @xmath23 such that @xmath95 , @xmath96 . @xmath18 conditionally satisfies a set of aggregates @xmath15 w.r.t . @xmath19 , if @xmath97 for every @xmath98 . given two sets @xmath18 and @xmath19 , and an aggregate program @xmath33 , the operator @xmath99 is defined as : @xmath100 . @xmath101 is monotone w.r.t . its first argument , given that the second argument is fixed . following @xcite , given a set of atoms @xmath11 , the least fixpoint of @xmath101 w.r.t @xmath11 is defined as @xmath102 , where @xmath103 and @xmath104 , for all @xmath105 . [ @xcite ] [ aggr - ans ] let @xmath33 be an aggregate program and @xmath11 a set of atoms . @xmath11 is an _ answer set _ of @xmath33 iff @xmath11 is a model of @xmath33 and @xmath106 . to present our results , it is notationally important to lift the concepts of conditional satisfaction and answer set to weight constraints . given a weight constraint @xmath7 , the _ domain of @xmath7 _ , denoted @xmath107 , is the set @xmath108 . let @xmath7 be a weight constraint and @xmath18 and @xmath19 be two sets of atoms . @xmath18 conditionally satisfies @xmath7 , w.r.t . @xmath19 , denoted @xmath109 , if for all @xmath23 such that @xmath110 , we have @xmath111 . first , the answer sets of a basic weight constraint program are defined using the concept of conditional satisfaction . [ answer - set ] let @xmath33 be a basic weight constraint program and @xmath11 a set of atoms . @xmath11 is an answer set of @xmath33 iff @xmath11 is a model of @xmath33 and @xmath112 . then , following @xcite , the answer sets of a general weight constraint program are defined as the answer sets of its _ instances_. let @xmath33 be a weight constraint program , @xmath28 a rule in @xmath33 of the form ( [ w - rule - form ] ) , and @xmath11 a set of atoms . the _ instance of @xmath28 w.r.t . @xmath11 _ is @xmath113 the _ instance of @xmath33 w.r.t . , denoted @xmath114 , is the program @xmath115 note that an instance of a program is a basic program . [ w - program - ans ] let @xmath33 be a weight constraint program and @xmath11 a set of atoms . @xmath11 is an answer set of @xmath33 iff @xmath11 is an answer set of the instance of @xmath33 w.r.t . @xmath11 . in the next section , we will show that , for some weight constraint programs the stable model and the answer set semantics coincide , while , for some others these semantics are different . before ending this section , we give a useful proposition , which shows a one - to - one correspondence between the stable models / answer sets of an arbitrary weight constraint program and those of its basic program counterparts . this result will be used later in this paper . [ basic ] let @xmath33 be a weight constraint program and @xmath11 a model of @xmath33 . define @xmath116 , then , @xmath11 is a stable model ( resp . answer set ) of @xmath33 iff @xmath11 is a stable model ( resp . answer set ) of @xmath117 . the correspondence between answer sets follows from definition [ w - program - ans ] above . for the correspondence between stable models , note that @xmath118 . in this section , we relate answer sets with stable models . first , we give a sufficient condition under which they agree with each other . then , we show the difference between these semantics , that is , stable models that are not answer sets may be circular justified , based on a formal notion of circular justification . at the end , we discuss the related justifications in the literature . we show that for a broad class of weight constraint programs , the stable models are precisely answer sets , and vice versa . given a weight constraint @xmath7 of the form ( [ w - form ] ) and a set of atoms @xmath11 , we define @xmath119 and @xmath120 . since @xmath7 is always clear by context , we will simply write @xmath121 and @xmath122 . let @xmath11 be a set of atoms and @xmath7 a weight constraint of the form ( [ w - form ] ) . @xmath7 is said to be _ strongly satisfiable by @xmath11 _ if @xmath16 implies that for any @xmath123 , @xmath124 . @xmath7 is _ strongly satisfiable _ if for any set of atoms @xmath11 , @xmath7 is strongly satisfiable by @xmath11 . a weight constraint program @xmath33 is _ strongly satisfiable _ if every weight constraint that appears in the body of a rule in @xmath33 is strongly satisfiable . intuitively , a strongly satisfiable weight constraint is a weight constraint whose upper bound is large enough to guarantee that , if a set of atoms satisfies the constraint , then any of its subset also satisfies the constraint . strongly satisfiable programs constitute a nontrivial class of programs . in particular , weight constraints @xmath7 that possess one of the following syntactically checkable conditions are strongly satisfiable . * @xmath8 contains only atoms ; * @xmath125 . note that upper - bound free weight constraints satisfy the second condition above . the following constraints are all strongly satisfiable : @xmath126~2 \\ 1~[a=1 , { \texttt{not } } b = 2]~3 \\ 1~[a=1,{\texttt{not } } b = 2 ] \end{array}\ ] ] but the weight constraint @xmath127~2\ ] ] is not , since it is satisfied by @xmath128 but not by @xmath129 . strongly satisfiable weight constraints are not necessarily convex or monotone . [ strong - satisfiable - not - convex ] let @xmath73 be the following weight constraint @xmath130\ ] ] since @xmath73 is upper bound free , it is strongly satisfiable . but @xmath73 is neither monotone nor convex , as @xmath131 , @xmath132 , and @xmath133 . we show that the stable model semantics coincides with the answer set semantics for strongly satisfiable programs . we need a lemma . [ key - lemma ] let @xmath7 be a weight constraint of the form ( [ w - form ] ) , and @xmath19 and @xmath11 be sets of atoms such that @xmath134 . then , * if @xmath135 then @xmath136 and @xmath137 . * if @xmath136 and @xmath7 is strongly satisfiable by @xmath11 , then @xmath135 . \(i ) we prove it by contraposition . that is , we show that if @xmath138 or @xmath139 , then @xmath140 . the case of @xmath138 is simple , which leads to @xmath141 hence @xmath142 . assume @xmath139 . by definition , the lower bound is violated , i.e. , @xmath143 , where @xmath144 . let @xmath145 , where @xmath146 and @xmath147 . since @xmath148 and @xmath143 , we have @xmath149 . then , from @xmath147 and the assumption @xmath139 , we get @xmath150 . it then follows from @xmath146 that @xmath151 . by construction , we have @xmath152 , and therefore we conclude @xmath142 . \(ii ) assume @xmath153 and @xmath7 is strongly satisfiable by @xmath11 . we show @xmath139 . we have either @xmath154 or @xmath155 . if @xmath155 then clearly @xmath139 . assume @xmath154 . then from @xmath142 , we have @xmath156 , @xmath157 , such that @xmath151 . since @xmath7 is strongly satisfiable by @xmath11 , if @xmath16 then for any @xmath158 , where @xmath123 , @xmath159 . assume @xmath16 . let @xmath18 be such that @xmath160 and @xmath161 . it is clear that @xmath159 leads to @xmath162 . thus , since @xmath16 , that @xmath151 is due to the violation of the lower bound , i.e. , @xmath163 . now consider @xmath164 , i.e. , we restrict @xmath165 to @xmath166 and expand @xmath167 to @xmath122 . note that by construction , it still holds that @xmath168 . clearly , that @xmath151 leads to @xmath169 , which is due to the violation of the lower bound , as @xmath170 , i.e. , we have @xmath171 . by definition , we have @xmath172 , where @xmath173 . note that since @xmath174 , we have @xmath175 . since @xmath176 , it follows that @xmath177 . now since @xmath178 is precisely the same constraint as @xmath51 , we have @xmath179 , and therefore @xmath143 . this shows @xmath139 . [ smodels - son ] let @xmath33 be a weight constraint program and @xmath56 . suppose for any weight constraint @xmath7 appearing in the body of a rule in @xmath33 , @xmath7 is strongly satisfiable by @xmath11 . then , @xmath11 is a stable model of @xmath33 iff @xmath11 is an answer set of @xmath33 . due to proposition [ basic ] , we only need to prove the claim for basic weight constraint programs . assume @xmath33 is a basic weight constraint program and @xmath11 a model of @xmath33 such that all weight constraints in @xmath33 are strongly satisfiable by @xmath11 . it suffices to prove that for any positive integer @xmath180 , @xmath181 , by induction on @xmath180 . _ base case _ : @xmath182 . we have @xmath183 . _ induction step _ : assume , for any @xmath184 , @xmath181 , and prove that @xmath185 . let @xmath35 be an atom such that @xmath186 and @xmath187 . then there exists a rule @xmath188 such that @xmath189 and @xmath190 , for each @xmath191 . it then follows from part ( ii ) of lemma [ key - lemma ] that @xmath192 . then @xmath193 by the induction hypothesis . so , @xmath194 . thus @xmath195 . similarly , we can show @xmath196 using part ( i ) of lemma [ key - lemma ] . thus @xmath185 . we therefore conclude @xmath197 . the following theorem follows from theorem [ smodels - son ] and the definition of strongly satisfiable programs . [ ss - program - answer - set ] let @xmath33 be a strongly satisfiable weight constraint program and @xmath56 . @xmath11 is a stable model of @xmath33 iff @xmath11 is an answer set of @xmath33 . [ coincide ] it has been shown that for logic programs with arbitrary abstract constraints the semantics based on conditional satisfaction are the most conservative in that the answer sets under this semantics are answer sets / stable models of a number of other semantics @xcite . it is then expected that the same holds true for programs with concrete constraints such as weight constraints . let @xmath33 be a weight constraint program . every answer set of @xmath33 is a stable model of @xmath33 , but the converse does not hold . let @xmath11 be a set of atoms . using part ( i ) of lemma [ key - lemma ] , it is easy to show by induction that for any positive integer @xmath180 , we have @xmath198 . for the converse , see the counterexample in example [ key - example ] below . question arises as why some stable models are not answer sets . later in section [ nested - expression ] , we will give a technical answer to this question . here , we suggest that in these extra stable models there may exist circular justifications . consider the following example . [ key - example ] let @xmath33 be a single - rule program : @xmath199~ 0\end{aligned}\ ] ] let @xmath200 and @xmath201 . the weight constraint @xmath202~ 0 $ ] in @xmath33 is not strongly satisfiable , since although @xmath203 satisfies the upper bound , its subset @xmath204 does not . both @xmath204 and @xmath203 are stable models by definition [ smodels - semantics ] . note that this is because @xmath205 and @xmath206 . but , @xmath204 is an answer set and @xmath203 is not , by definition [ w - program - ans ] . the reason that @xmath203 is not an answer set of @xmath33 is due to the fact that @xmath35 is derived by its being in @xmath203 . this kind of circular justification can be seen more intuitively below using equivalence substitutions . * the weight constraint is substituted with an equivalent aggregate : @xmath207 where @xmath208 . * the weight constraint is transformed to an equivalent one without negative literal , but with a negative weight , according to @xcite : @xmath209{-1}\ ] ] * the weight constraint is substituted with an equivalent _ abstract constraint atom _ @xcite , where @xmath210 is a finite set of ground atoms called the _ domain _ , and @xmath211 is a collection of subsets of @xmath210 called _ admissible solutions_. in this example , the set @xmath212 satisfies the abstract constraint atom , since the admissible solution @xmath129 in it is satisfied by @xmath23 . ] : @xmath213 for the claim of equivalence , note that for any set of atoms @xmath11 , we have : @xmath2140 $ ] iff @xmath215{-1}$ ] iff @xmath216 iff @xmath217 . for logic programs with abstract constraint atoms , it is often said that all of the major semantics coincide for programs with monotone constraints . for example , this is the case for the semantics proposed in @xcite . what is unexpected is that this is not the case for the stable model semantics for weight constraint programs . by the standard definition of monotonicity , the constraint @xmath202~ 0 $ ] is actually monotone ! one may think that the culprit for @xmath203 above is because it is not a minimal model . however , the following example shows that stable models that are minimal models may still be circularly justified . consider the following weight constraint program @xmath33 ( obtained from the one in example [ key - example ] by adding the second rule ) : @xmath199~ 0\\ f \leftarrow { \texttt{not } } f , { \texttt{not } } a\end{aligned}\ ] ] now , @xmath218 is a minimal model of @xmath33 , and also a stable model of @xmath33 , but clearly @xmath35 is justified by its being in @xmath11 . we now give a more formal account of _ circular justification _ for stable models , borrowing the idea of _ unfounded sets _ previously used for normal programs @xcite and logic programs with monotone and antimonotone aggregates @xcite . [ cj ] let @xmath33 be an weight constraint program and @xmath11 a stable model of @xmath33 . @xmath11 is said to be _ circularly justified _ , or simply _ circular _ , if there exists a non - empty set @xmath219 such that @xmath220 , @xmath221 does not satisfy the body of any rule @xmath188 such that @xmath222 . otherwise @xmath11 is said to be _ non - circular_. [ as - not - cj ] let @xmath33 be a weight constraint program and @xmath11 a stable model of @xmath33 . if @xmath11 is an answer set of @xmath33 , then @xmath11 is not circular . let @xmath33 be a weight constraint program and @xmath11 an answer set of @xmath33 . assume @xmath11 is circular . then there exists a non - empty subset @xmath223 such that @xmath220 , @xmath224 does not satisfy the body of any rule @xmath188 such that @xmath222 . by a simple induction on the construction of @xmath225 , it can be shown that for each of such @xmath226 , we have @xmath227 . this contradicts the assumption that @xmath11 is an answer set . example [ key - example ] shows that extra stable models ( the stable models that are not answer sets ) of a program may be circular . however , not all extra stable models are necessarily circular , according to definition [ cj ] . therefore , the notion of circularity given in definition [ cj ] only serves as a partial characterization of circular justification . [ non - cj - extra ] consider a weight constraint program @xmath33 that consists of the following three rules . @xmath228 \label{rule-2 } \\ b & \leftarrow & [ { \texttt{not } } b=1]0 \label{rule-3}\end{aligned}\ ] ] @xmath229 is a stable model but not an answer set of @xmath33 . however , it can be verified that @xmath11 is not circular under definition [ cj ] : @xmath230 can be derived from the first rule if we do nt have @xmath230 , and by the second rule if we do . we shall comment that other forms of non - circular nature of answer sets have been formulated in different ways , e.g. , by the existence of a level mapping @xcite and by a translation of a constraint to sets of solutions @xcite . a semantics is a formal account of intuitions of what justifications for atoms in a stable model ought to be . for weight constraint programs , there are different intuitions . consider the single - rule program @xmath33 in example [ key - example ] again @xmath2310\end{aligned}\ ] ] one possible interpretation of how @xmath35 is `` justified '' to be in a stable model is by using the transformation proposed in @xcite . by the transformation , @xmath33 is translated to @xmath232 , which consists the following rules : @xmath2330\\ b & \leftarrow & 0[a=1]0\end{aligned}\ ] ] it can be verified that the set @xmath234 is a stable model of @xmath232 . the justification of @xmath35 is : given @xmath11 , @xmath230 can not be in any stable model of @xmath232 by the second rule , then @xmath35 can be derived by the first rule . note that the above transformation introduces the new atom @xmath230 in the translated program @xmath232 . it assumes that @xmath235 implies something new to the original program . the justification of atom @xmath35 depends on the truth status of the new atom @xmath230 . whether such a justification is intuitive seems arguable . another justification of @xmath35 is by transforming the program @xmath33 to a program with nested expressions @xcite . we will discuss this in more details later in section [ relation - fl - translation ] . for this example , we also note that the set @xmath234 is not an answer set under any semantics based on computations studied in @xcite . we show that the gap between the answer set semantics and the stable model semantics can be closed by a transformation , which translates a weight constraint program to a strongly satisfiable program whose stable models are free of circular justifications . in this way , we are able to apply current asp systems that implement the stable model semantics to compute answer sets for weight constraint programs . in particular , later on we will use strongly satisfiable programs to represent aggregate programs , so that an implementation of aggregate programs can be realized by an implementation of weight constraint programs . we present an encoding of a weight constraint , where a weight constraint is represented by two strongly satisfiable weight constraints . this encoding captures conditional satisfaction for weight constraints in terms of standard satisfaction . in other words , for weight constraints , the encoding allows conditional satisfaction to be checked by standard satisfaction . let @xmath7 be a weight constraint of the form ( [ w - form ] ) . the _ strongly satisfiable encoding _ of @xmath7 , denoted @xmath236 , consists of the following constraints : @xmath237 \\ \\ w_u : -u+ \sum_{i=1}^{n}{w_{a_i}}+ \sum_{i=1}^{m}{w_{b_i}}[{\texttt{not } } a_1=w_{a_1 } , ... , { \texttt{not } } a_n = w_{a_n } , b_1=w_{b_1 } , ... , b_m = w_{b_m } ] \end{array}\ ] ] intuitively , @xmath238 and @xmath239 are to code the lower and upper bound constraints of @xmath7 , respectively . it is easy to verify that the encoding is satisfaction - preserving , as shown below . [ sat - preserving ] let @xmath7 be a weight constraint , @xmath236 be its strongly satisfiable encoding , and @xmath11 be a set of atoms . @xmath12 iff @xmath240 and @xmath241 . the satisfaction of @xmath238 is trivial , since @xmath238 is just the lower bound part of @xmath7 . we show that @xmath239 is the upper bound part of @xmath7 . note that the upper bound part of @xmath7 is @xmath242 which is equivalent to @xmath243 by the transformation that eliminates the negative weights ( introduced in section [ sec - sm - semantics ] ) , the constraint ( [ c-2 ] ) is equivalent to the weight constraint @xmath239 . using lemmas [ key - lemma ] and [ sat - preserving ] , we establish the following theorem . [ cond - sat - reduct ] let @xmath7 be a weight constraint , @xmath236 be the strongly satisfiable encoding of @xmath7 , and @xmath19 and @xmath11 be two sets of atoms such that @xmath244 . @xmath245 iff @xmath246 and @xmath247 . ( @xmath248 ) since @xmath249 is the same as @xmath51 , by part ( i ) of lemma [ key - lemma ] , we have @xmath250 . in the following , we show @xmath247 . assume @xmath245 and @xmath244 . then , by definition , we have @xmath22 , and @xmath251 such that @xmath252 and @xmath253 , @xmath111 . let @xmath254 such that @xmath255 and @xmath256 . under this notation , from the assumption @xmath245 and @xmath244 , we get @xmath252 and @xmath257 . it follows that @xmath162 , that is , @xmath258 . this implies @xmath259 , from which the following inequations can be derived @xmath260 the last one shows @xmath247 . ( @xmath261 ) assume @xmath250 and @xmath262 . since neither has an upper bound , both of them are strongly satisfiable . from part ( ii ) of lemma [ key - lemma ] , we have @xmath263 and @xmath264 . it then follows @xmath251 such that @xmath265 and @xmath257 , @xmath266 and @xmath267 . then by theorem [ sat - preserving ] , we have @xmath251 such that @xmath265 and @xmath257 , @xmath111 . this shows @xmath245 . using the strongly satisfiable encoding of weight constraints , a weight constraint program can be translated to a strongly satisfiable program , so that the answer sets of the original program are precisely the stable models of the translated program and vice versa .. let @xmath33 be a weight constraint program . the _ strongly satisfiable translation _ of @xmath33 , denoted @xmath268 , is the program obtained by replacing each @xmath7 in the body of rules in @xmath33 by the strongly satisfiable encoding of @xmath7 . [ ans - strong - programs ] let @xmath33 be a weight constraint program and @xmath11 a set of atoms . @xmath11 is an answer set of @xmath33 iff @xmath11 is a stable model of @xmath268 . due to proposition [ basic ] , we only need to prove the claim for basic weight constraint programs . using theorem [ cond - sat - reduct ] , we have a one - to - one correspondence between the derivations based on conditional satisfaction ( definition [ answer - set ] ) and the derivations in the construction of the least model ( definition [ sm - fixpoint ] ) , which can be shown by an easy induction on the length of these constructions . consider a program @xmath33 with a single rule : @xmath269 2\ ] ] @xmath268 consists of @xmath269,~1 [ a = 3]\ ] ] the weight constraints in @xmath268 are all upper bound - free , hence @xmath268 is strongly satisfiable . both @xmath270 and @xmath129 are stable models of @xmath33 , but @xmath270 is the only stable model of @xmath268 , which is also the only answer set of @xmath33 . in this section , we propose an approach to computing the answer sets of an aggregate program . for this , we translate an aggregate program to a strongly satisfiable weight constraint program and then compute its stable models as answer sets . this section shows that the aggregates can be encoded as weight constraints . in the following , given sets @xmath11 and @xmath19 , it is convenient to express the set @xmath11 , restricted to @xmath19 , as @xmath271 which is defined by @xmath272 . [ w - encoding ] let @xmath73 be an aggregate in the form ( [ aggregate - form ] ) . a set of weight constraints @xmath273 is an _ weight constraint encoding _ ( or _ encoding _ ) of @xmath73 , denoted @xmath274 , if for any model @xmath11 of @xmath73 , there is a model @xmath275 of @xmath274 such that @xmath276 , and for any model @xmath275 of @xmath274 , @xmath277 is a model of @xmath73 . we show the encoding of aggregates of the form ( [ aggregate - form ] ) , where the operator @xmath278 is @xmath279 . the encoding can be easily extended to other relational operators except for the operator @xmath280 ( more on @xmath280 later in this section ) . for example , aggregate @xmath281 can be expressed as @xmath282 . the encoding works for the aggregates whose variable domain contains only integers . for the aggregates whose variable domain contains real numbers , each real number can be converted to an integer by multiplying a factor . in this case , the @xmath72 ( in the formula ( [ aggregate - form ] ) ) also needs to be processed accordingly . for convenience , below we may write negative weights in weight constraints . recall that negative weights can be eliminated by a simple transformation . * @xmath283 * + these aggregates can be encoded by weight constraints rather directly . for instance , aggregate @xmath284 can be represented by @xmath285\end{aligned}\ ] ] where the domain of the aggregate is @xmath286 . note that , a multiset in @xmath284 can be encoded directly by a weight constraint , since the latter does not require distinct literals in it . for example , for the aggregate @xmath287 where @xmath68 is defined by the multiset @xmath288 , the aggregate can be encoded by the weight constraint @xmath289 $ ] . we note that aggregates @xmath290 and @xmath291 can be encoded simply by substituting the weights in ( [ sum ] ) with @xmath292 and @xmath293 ( for avg the lower bound @xmath180 is also replaced by zero ) , respectively . + * @xmath294 * + let @xmath295 be an aggregate . the idea in the encoding of @xmath73 is that for a set of numbers @xmath296 , the maximum number in @xmath19 is greater than or equal to @xmath180 if and only if @xmath297 for each atom @xmath298 , two new literals @xmath299 and @xmath300 are introduced . the encoding @xmath274 consists of the following constraints . i \leq n\\ \label{max-2 } 0~[p(a_i)=-d_i , p^{+}(a_i)=d_i],~1\leq i \leq n\\ \label{max-3 } 0~[p(a_i)=d_i , p^{-}(a_i)=-d_i],~1\leq i \leq n\\ \label{max-4 } 1~[p(a_1)=d_1 , p^{+}(a_1)=d_1 , p^{-}(a_1)=-d_1 , \nonumber \\ ... , p(a_n)=d_n , p^{+}(a_n)=d_n , p^{-}(a_n)=-d_n]\\ \label{max-5 } 1~[p(a_1)=1 , ... , p(a_n)=1]\end{aligned}\ ] ] where @xmath302 . in the following , for any model @xmath11 of such an encoding , @xmath303 means @xmath304 and @xmath305 means @xmath306 . the constraints ( [ max-1 ] ) , ( [ max-2 ] ) and ( [ max-3 ] ) are used to encode @xmath307 . clearly , if @xmath308 , we have @xmath309 and @xmath310 ; if @xmath311 , we have @xmath312 and @xmath313 ; and if @xmath314 , we have @xmath309 or @xmath312 . the constraint ( [ max-4 ] ) encodes the relation ( [ max - relation ] ) and the constraint ( [ max-5 ] ) guarantees that a model of @xmath274 is not an empty set . + * @xmath315 * + let @xmath316 be an aggregate . the idea in the encoding of @xmath73 is that for a set of numbers @xmath296 , the minimal number in @xmath19 is greater than or equal to @xmath180 if and only if @xmath317 similar to @xmath294 , the aggregate @xmath315 can be encoded by the following weight constraints . i \leq n\\ \label{min-2 } 0~[p^{+}(a_i)=d_i , p(a_i)=-d_i],~1\leq i \leq n\\ \label{min-3 } 0~[p^{-}(a_i)=-d_i , p(a_i)=d_i],~1\leq i \leq n\\ \label{min-4 } 0~[p(a_1)=d_1 , p^{+}(a_1)=-d_1 , p^{-}(a_1)=d_1 , \nonumber \\ ... , p(a_n)=d_n , p^{+}(a_n)=-d_n , p^{-}(a_n)=d_n]~ 0\\ \label{min-5 } 1~[p(a_1)=1 , ... , p(a_n)=1]\end{aligned}\ ] ] where @xmath319 . the constraint ( [ min-1 ] ) , ( [ min-2 ] ) and ( [ min-3 ] ) are the same to the first three constraints in the encoding of @xmath294 ( except for the value of @xmath320 ) , respectively . the constraint ( [ min-4 ] ) encodes the relation ( [ min - relation ] ) and the constraint ( [ min-5 ] ) guarantees that a model of @xmath274 is not an empty set . + we note that all the encodings above result in weight constraints whose collective size is linear in the size of the domain of the aggregate being encoded . in the encoding of @xmath294 ( similarly for @xmath315 ) , the first three constraints are the ones between the literal @xmath298 and the newly introduced literals @xmath299 and @xmath300 . we call them _ auxiliary constraints_. the last two constraints code the relation between @xmath298 and @xmath321 , where @xmath322 . we call them _ relation constraints_. let @xmath73 be an aggregate , we denote the set of auxiliary constraints in @xmath274 by @xmath323 and the set of relation constraints by @xmath324 . if @xmath73 is aggregate @xmath325 , @xmath326 , or @xmath327 , we have that @xmath328 , because no new literals are introduced in their encodings . [ encoding - aggr ] the set of weight constraint ( [ sum ] ) , the set of weight constraints from ( [ max-1 ] ) to ( [ max-5 ] ) , and the set of weight constraints from ( [ min-1 ] ) to ( [ min-5 ] ) , are weight constraint encodings ( definition [ w - encoding ] ) of the aggregates @xmath325 , @xmath294 , and @xmath315 , respectively . the proof for the encoding of aggregate @xmath325 is straightforward . the proof for the encoding of aggregate @xmath315 is similar to that for @xmath294 , which we show below . let @xmath11 be a set of atoms and @xmath78 . suppose @xmath329 and @xmath330 . then , we can construct @xmath275 as follows : a. @xmath331 and @xmath332 , if @xmath333 and @xmath334 ; b. @xmath331 and @xmath335 , if @xmath333 and @xmath336 . we use @xmath337 to @xmath338 to denote the weight constraints in ( [ max-1 ] ) to ( [ max-5 ] ) . it is easy to check that the weight constraints @xmath337 , @xmath339 , and @xmath340 are satisfied by @xmath275 . since @xmath330 , we have @xmath341 and @xmath342 . therefore @xmath343 and @xmath338 are also satisfied by @xmath275 . so @xmath344 . let @xmath275 be a set of atoms and @xmath344 . since @xmath275 satisfies @xmath337 , @xmath339 and @xmath340 , we have @xmath309 and @xmath310 , for @xmath334 ; @xmath312 and @xmath313 , for @xmath311 ; and @xmath309 or @xmath312 , if @xmath314 . since @xmath345 and @xmath346 , there must be an @xmath347 , such that @xmath334 and @xmath348 . that is , @xmath349 . then , we have @xmath78 . we translate an aggregate program @xmath33 to a weight constraint program , denoted @xmath350 , as follows : 1 . for each rule of the form ( [ aggr - rule - form ] ) in @xmath33 , we include in @xmath351 a weight rule of the form @xmath352 where @xmath353 is the conjunction of the weight constraints that encode the aggregate @xmath354 ; and 2 . if there are newly introduced literals in the encoding of aggregates , the _ auxiliary rule _ of the form @xmath355 is included in @xmath351 , for each auxiliary constraint @xmath7 of each atom @xmath298 in the aggregates . note that a weight constraint program @xmath33 can be translated to a strongly satisfiable program using the translation @xmath268 given in section [ transform ] . we have the following theorem establishing the correctness of the transformation @xmath356 . [ thm - translation ] let @xmath33 be an aggregate program where the relational operator is not @xmath280 . for any stable model @xmath11 of @xmath357 , @xmath358 is an answer set of @xmath33 . for any answer set @xmath11 for @xmath33 , there is a stable model @xmath275 of @xmath357 such that @xmath359 . the rules of the form ( [ transformed - rule - form ] ) are the translated counterpart of the rules in @xmath33 . the auxiliary rules of the form ( [ auxiliary - rule - form ] ) are added to enforce the auxiliary constraints . note that @xmath357 is a strongly satisfiable program . then the theorem follows from theorem [ ans - strong - programs ] and theorem [ encoding - aggr ] . * remark * for an aggregate where the relation operator is not @xmath280 , the aggregate can be encoded by a conjunction of weight constraints as we have shown in this section . in this case , logic equivalence leads to equivalence under conditional satisfaction . that is why we only need to ensure that an encoding is satisfaction - preserving . for an aggregate where the relation operator is @xmath280 , two classes are distinguished . one consists of aggregates of the forms @xmath360 . for these aggregates , the operator @xmath280 can be treated as the disjunction of the operators @xmath361 and @xmath362. consider the aggregate @xmath363 . @xmath73 is logically equivalent to @xmath364 , where @xmath365 and @xmath366 . let @xmath18 and @xmath19 be two sets of atoms , it is easy to show that @xmath93 iff @xmath367 or @xmath368 . the other class consists of the aggregates of the forms @xmath369 , @xmath370 , @xmath371 , and @xmath372 . for these aggregates , the operator @xmath280 can not be treated as the disjunction of @xmath361 and @xmath362 , since the conditional satisfaction may not be preserved . below is an example . consider the aggregates @xmath373 , @xmath374 , and @xmath375 . note that @xmath73 is logically equivalent to @xmath364 . consider @xmath376 and @xmath377 . while @xmath19 conditionally satisfies @xmath73 w.r.t . @xmath11 ( i.e. , @xmath378 ) , it is not the case that @xmath19 conditionally satisfies @xmath379 w.r.t . @xmath11 or @xmath19 conditionally satisfies @xmath380 w.r.t . @xmath11 . in this section , we further relate answer sets with stable models in terms of logic programs with nested expressions . we formulate a transformation of weight constraint programs to programs with nested expressions and compare this transformation to the one in @xcite . the comparison reveals that the difference of the semantics lies in the different interpretations of the constraint on the upper bounds of weight constraints in a program : while our transformation interprets it directly , namely as `` less than or equal to '' , the one in @xcite interprets it as `` not greater than '' , which may create double negations ( the atoms that are preceded by ` not ` ` not ` ) in nested expressions . it is the semantics of these double negations that differentiates the two semantics . in the language of nested expressions , _ elementary formulas _ are atoms . the classical negation is irrelevant here . ] and symbols @xmath381 ( false ) and @xmath382 ( true ) . _ formulas _ are built from elementary formulas using the unary connective @xmath383 and the binary connectives , ( conjunction ) and ; ( disjunction ) . a rule with nested expressions is of the form @xmath384 where both @xmath385 and @xmath386 are formulas . for a rule @xmath28 of the form ( [ nested - rule - form ] ) , we use @xmath29 and @xmath31 to denote the @xmath386 and the @xmath385 of @xmath28 , respectively . a program with nested expressions is a set of rules with nested expressions . the satisfaction of a formula by a set of atoms @xmath11 is defined as follows : * for a literal @xmath1 , @xmath387 if @xmath388 * @xmath389 * @xmath390 * @xmath391 if @xmath392 and @xmath393 * @xmath394 if @xmath392 or @xmath393 * @xmath395 if @xmath396 . the reduct of a formula @xmath397 with respect to a set of atoms @xmath11 , denoted @xmath398 , is defined recursively as follows : * for an elementary formula @xmath397 , @xmath399 * @xmath400 * @xmath401 * @xmath402 the reduct of a program @xmath33 with respect to a set of atoms @xmath11 is the set of rules @xmath403 for each rule of the form ( [ nested - rule - form ] ) in @xmath33 . the concept of a stable model is defined as follows . [ @xcite ] let @xmath33 be a logic program with nested expressions and @xmath11 a set of atoms . @xmath11 is a stable model of @xmath33 if @xmath11 is a minimal model of @xmath54 . we present a nested expression encoding , called the _ direct nested expression encoding _ of weight constraints . we show that conditional satisfaction of a weight constraint can be captured by the standard satisfaction of the reduct of the encoding of the weight constraint . in the rest of the paper , we will use the following notation : for a set of literals @xmath68 , we define @xmath404 and @xmath405 . given a weight constraint @xmath7 of the form ( [ w - form ] ) , the _ nested expression encoding _ of @xmath7 , denoted @xmath406 , is the formula @xmath407\end{aligned}\ ] ] where @xmath408 is the set of atoms in @xmath7 . for @xmath409 and @xmath410 for @xmath411 . then we can rewrite this formula as @xmath412\ ] ] ] intuitively , @xmath406 is a nested expression representing the sets that satisfy @xmath7 . let @xmath4131 $ ] be a weight constraint . the subsets of @xmath408 that satisfy @xmath7 are @xmath414 and @xmath415 . thus @xmath416 . an interesting result of the directed encoding is that for the resulting nested expression , conditional satisfaction is precisely the satisfaction of the _ reduct _ of the expression . that is , given a weight constraint @xmath7 and two sets of atoms @xmath134 , @xmath417 if and only if @xmath418 . before giving this result , we prove a lemma first . [ nested - lemma-1 ] let @xmath7 be a weight constraint , @xmath406 its nested expression encoding , and @xmath236 its strongly satisfiable encoding . then , for any two sets of atoms @xmath19 and @xmath11 such that @xmath244 , @xmath419 iff @xmath246 and @xmath247 . in this proof , given a weight constraint @xmath7 of the form ( [ w - form ] ) , we denote @xmath420 and @xmath421 ; we express @xmath406 as a disjunction of conjunctions @xmath422 s , i.e. , @xmath423 , for some @xmath424 . for notational convenience , such a conjunction may be referred to as a set as well , i.e. , given a conjunction @xmath425 , we may use the same @xmath422 to denote the set @xmath426 , and vice versa . given a set @xmath19 , we use the notation @xmath427 . since the proof is mainly about mathematical transformation , for convenience , we present it as a set of mechanical inferences . ( @xmath248 ) we give a detailed proof for @xmath250 . the proof of @xmath247 is similar . \(1 ) assume @xmath418 and @xmath134 . \(2 ) @xmath428 such that @xmath429 , due to ( 1 ) and the definition of @xmath406 . \(3 ) @xmath430 , we have @xmath431 , that is , @xmath432 , due to ( 2 ) . \(4 ) @xmath433 , due to ( 2 ) . \(5 ) @xmath434 , due to ( 2 ) . \(6 ) @xmath435 , due to ( 5 ) . \(7 ) @xmath436 , due to ( 4 ) . \(8 ) @xmath437 where @xmath438 , due to ( 3 ) . \(9 ) @xmath439 , due to ( 6 ) , ( 7 ) and ( 8) . \(10 ) @xmath440 , due to ( 9 ) . \(11 ) @xmath441 , due to ( 10 ) . \(12 ) @xmath246 , due to ( 11 ) . ( @xmath261 ) \(1 ) assume @xmath246 , @xmath247 and @xmath134 . \(2 ) @xmath442 , due to ( 1 ) . \(3 ) @xmath443 , due to ( 1 ) . \(4 ) @xmath444 , due to ( 3 ) . \(5 ) let @xmath445 . \(6 ) @xmath446 , due to ( 5 ) and ( 1 ) . \(7 ) @xmath447 , due to ( 4 ) and ( 6 ) . \(8 ) @xmath448 , due to ( 2 ) and ( 7 ) . \(9 ) @xmath449 , due to ( 8) . \(10 ) @xmath450 , due to ( 9 ) . \(11 ) @xmath451 , due to ( 5 ) . \(12 ) @xmath418 , due to ( 11 ) . [ nested - theorem-1 ] let @xmath7 be a weight constraint , and @xmath19 and @xmath11 two sets of atoms . then , @xmath245 iff @xmath419 . this follows from theorem [ cond - sat - reduct ] and lemma [ nested - lemma-1 ] . using the direct nested expression encoding of weight constraints , a weight constraint program can be translated to a program with nested expressions , such that the answer sets of the original program are precisely the stable models of the translated program and vice versa . let @xmath33 be a weight constraint program and @xmath28 a rule of the form ( [ w - rule - form ] ) in @xmath33 . the _ nested expression translation of @xmath28 _ , denoted @xmath452 , is the rule of the form @xmath453 where @xmath454 are the positive literals in @xmath30 . intuitively , the conjunctive term @xmath455 represent that we are free to choose the atoms in the head of the rule to include in an answer set . let @xmath33 be a weight constraint program . the _ nested expression translation of @xmath33 _ , denoted @xmath456 , is the program obtained by replacing each rule @xmath28 in @xmath33 by @xmath452 . let @xmath33 be a weight constraint program , @xmath28 a rule in @xmath33 , and @xmath11 a set of atoms . by the definitions of @xmath452 and the reduct of a nested expression , we know that the reduct of @xmath452 w.r.t . @xmath11 , denoted @xmath457 , is the rule of the form @xmath458 where @xmath459 . we will use this fact in the proofs below . our main result is that the answer sets of a weight constraint program coincide with the stable models of its nested expression translation . to establish this , we need to show a one - to - one correspondence between the least fixpoint of the operator @xmath460 applied on a weight constraint program is the unique minimal model of the reduct of its nested expression translation . to show the main theorem , we prove two lemmas firstly . one shows the coincidence of the satisfaction of a weight constraint , its nested expression encoding , and the reduct of the nested encoding . then using this lemma , we show the coincidence of the models of a weight constraint program , its nested expression translation , and the reduct of its nested expression translation . the later lemma helps to establish the correspondence between the least fixpoint of the operator @xmath460 and the minimal model of the program reduct . [ ne - reduct - satisfaction ] let @xmath7 be a weight constraint of the form ( [ w - form ] ) , @xmath406 its nested expression encoding , and @xmath11 a set atoms . then , @xmath16 iff @xmath461 iff @xmath462 . that @xmath12 iff @xmath461 follows directly from the definition of @xmath406 . we give a proof of the claim @xmath461 iff @xmath462 . suppose @xmath463 . ( @xmath248 ) since @xmath464 , there is an @xmath450 such that @xmath465 and @xmath466 . by the definition of the reduct of a nested expression , we have @xmath467 and then @xmath462 . ( @xmath261 ) since @xmath462 , there is an @xmath450 such that @xmath468 . by the definition of reduct , we have @xmath465 and @xmath466 . therefore @xmath469 and then @xmath461 . [ ne - program - reduct - model ] let @xmath33 a weight constraint program and @xmath11 a set of atoms . @xmath57 iff @xmath470 iff @xmath471 . the equivalence @xmath57 iff @xmath470 holds simply because @xmath456 is satisfaction - preserving . we give a proof of the statement @xmath470 iff @xmath471 . ( @xmath248 ) suppose @xmath470 . let @xmath472 be a rule of the form ( [ ne_rule_reduct ] ) . if @xmath473 for all @xmath474 , then by lemma [ ne - reduct - satisfaction ] , we have @xmath475 for all @xmath474 . since @xmath470 , then @xmath476 again by lemma [ ne - reduct - satisfaction ] , we know @xmath477 . since @xmath459 , we have @xmath478 . as @xmath479 is arbitrary in @xmath480 , we conclude @xmath471 . ( @xmath261 ) suppose @xmath471 . let @xmath481 be a rule of the form ( [ ne - rule - form ] ) . if @xmath475 for all @xmath482 , then by lemma [ ne - reduct - satisfaction ] , we have @xmath473 for all @xmath474 . since @xmath471 , then @xmath477 again by lemma [ ne - reduct - satisfaction ] , we know @xmath476 . it is obvious that @xmath483 . so @xmath484 . as @xmath28 is any rule in @xmath456 , we have @xmath470 . [ nested - theorem ] let @xmath33 be a weight constraint program , @xmath456 the nested expression translation of @xmath33 , and @xmath11 a set of atoms . then , @xmath11 is an answer set of @xmath33 iff @xmath11 is a stable model of @xmath456 . ( @xmath248 ) let @xmath33 be a weight constraint program and @xmath11 an answer set of @xmath33 . then @xmath57 and @xmath485 . by lemma [ ne - program - reduct - model ] , @xmath470 and @xmath471 . we prove by contradiction that @xmath11 is a minimal model of @xmath480 . suppose for some @xmath486 such that @xmath487 . note that by lemma [ ne - program - reduct - model ] , we also have @xmath488 . let @xmath489 . then , there exists a rule @xmath490 of the form ( [ w - rule - form ] ) satisfying that @xmath491 and @xmath492 such that @xmath493 , for all @xmath494 . note that @xmath495 , hence theorem [ nested - theorem-1 ] is applicable , from which we know @xmath496 , for all @xmath494 . since @xmath497 , we must have @xmath498 , which contradicts to the assumption that @xmath499 . we therefore conclude that @xmath11 is a minimal model of @xmath480 , i.e. , @xmath11 is a stable model of @xmath456 . ( @xmath261 ) let @xmath33 be a weight constraint program and suppose @xmath11 is a minimal model of @xmath480 . by lemma [ ne - program - reduct - model ] , @xmath11 is a model of @xmath33 . by the definitions of the operator @xmath101 and @xmath114 , we have @xmath500 . let @xmath501 . we will prove that @xmath502 . for this , let s assume it is not the case , i.e. , @xmath503 . since @xmath11 is a minimal model of @xmath480 , we have @xmath504 . then there is a rule @xmath28 of the form ( [ w - rule - form ] ) in @xmath33 and its corresponding rule @xmath452 of the form ( [ ne_rule_reduct ] ) in @xmath480 such that @xmath505 for all @xmath506 , and @xmath507 . since @xmath508 , for the rule @xmath452 , we have @xmath509 . it follows that for some @xmath510 , @xmath511 . as @xmath507 , it must be the case that @xmath512 such that @xmath513 . by theorem [ nested - theorem-1 ] , however , that @xmath505 for all @xmath506 leads to @xmath514 for all @xmath515 . then , by the definitions of the operator @xmath101 and @xmath114 , we must have @xmath516 . this is a contradiction . therefore , it must be the case that @xmath502 , and it follows that @xmath11 is an answer set of @xmath33 . consider the program @xmath33 in example [ non - cj - extra ] . its nested expression translation @xmath456 consists of the following rules . @xmath517 it can be verified that the only stable model of @xmath456 is @xmath129 , which is also the only answer set of @xmath33 . ferraris and lifschitz @xcite proposed a nested expression encoding of weight constraints . using this encoding , a weight constraint program can be translated to a program with nested expressions , such that there is a one - to - one correspondence between the stable models of the weight constraint program and the stable models of its translated program with nested expressions . the difference between our nested expression translation and ferraris and lifschitz s translation ( fl - translation ) lies in the interpretation of the upper bound constraint of a weight constraint . to illustrate this difference , let s denote a weight constraint @xmath7 of the form ( [ w - form ] ) by @xmath518u$ ] , where @xmath3=[a_1 = w_{a_1}, ... ,a_n = w_{a_n } , { \texttt{not } } b_1= w_{b_1}, ... ,{\texttt{not } } b_m= w_{b_m}]$ ] . we call that @xmath518 $ ] and @xmath3u$ ] the lower bound constraint and upper bound constraint of @xmath7 , respectively . obviously , the lower bound and upper bound constraints are also weight constraints . in our translation , the upper bound constraint @xmath3u$ ] is _ directly _ encoded by the sets of atoms that satisfy it . in the fl - translation , @xmath3u$ ] is encoded as @xmath519 $ ] , where @xmath520 $ ] is further encoded as the sets of atoms that satisfy the weight constraint @xmath520 $ ] , possibly creating double negations . this difference is the _ only _ reason that the stable models of our translated program are the answer sets of the original program while the stable models of the fl - translated program are the stable models of the original program . it should be clear that the extra stable models that are not answer sets are created by double negations generated by the indirect interpretation in the fl - translation . we use the following example for an illustration . consider the program @xmath33 in example [ key - example ] , which consist of a single rule @xmath5210\end{aligned}\ ] ] by our translation , @xmath456 consists of @xmath522 the only stable model of @xmath456 is @xmath270 , which is the unique answer set of @xmath33 . by the fl - translation , the translated program @xmath117 is @xmath523 the stable models of @xmath117 are @xmath270 and @xmath129 . among them , the set @xmath129 is not an answer set , but it is justified by the stable model semantics through the double negation @xmath524 . the theoretical studies show that an aggregate program can be translated to a weight constraint program whose stable models are precisely the answer sets of the original program . this leads to a prototype implementation called alparse to compute the answer sets for aggregate programs . in alparse , an aggregate program is firstly translated to a strongly satisfiable program using the translation given in section [ aggr - wprogram ] , then the stable models of the translated strongly satisfiable program are computed using an asp solver that implements the stable model semantics for weight constraint programs . in the next two subsections , we use smodels version 2.34 and clasp version 2.0.3 respectively as the underlying asp solver of alparse and compare alparse with the implementations of aggregate programs smodels@xmath525 and dlv version 2007 - 10 - 11 . the experiments are run on scientific linux release 5.1 with 3ghz cpu and 1 gb ram . the reported time of alparse consists of the transformation time ( from aggregate programs to strongly satisfiable programs ) , the grounding time ( calling to lparse version 1.1.2 for smodels and gringo version 2.0.3 for clasp ) , and the search time ( by smodels or clasp ) . the time of smodels@xmath525 consists of grounding time , search time and unfolding time ( computing the solutions to aggregates ) . the time of dlv includes the grounding time and search time ( the grounding phase is not separated from the search in dlv ) . all times are in seconds . in this section , we compare our approach with two systems , smodels@xmath525 and dlv . * comparison with smodels@xmath525 * we compare the encoding approach proposed in last section to the unfolding approach implemented in the system smodels@xmath525 @xcite.ielkaban / asp - aggr.html . ] the aggregates used in the first and second set of problems ( the company control and employee raise problems ) are @xmath325 ; the third set of problems ( the party invitation problems ) are @xmath326 , and the fourth and fifth set of problems ( the nm1 and nm2 , respectively ) are @xmath294 and @xmath315 , respectively . the experimental results are reported in table [ smodels - a ] , where the `` sample size '' is measured by the argument used to generate the test cases . the times are the average of one hundred randomly generated instances for each sample size . the results show that smodels is often faster than smodels@xmath525 , even though both use the same search engine . scale - up could be a problem for smodels@xmath525 , due to exponential blowup . for instance , for an aggregate like @xmath526 , smodels@xmath525 would list all _ aggregate solutions _ @xcite in the unfolded program , whose number is @xmath527 . for a large domain @xmath19 and @xmath180 being around @xmath528 , this is a huge number . if one or a few solutions are needed , alparse takes much less time to compute the corresponding weight constraints than smodels@xmath525 . * comparison with dlv * in @xcite the seating problem was chosen to evaluate the performance of dlv . the problem is to generate a sitting arrangement for a number of guests , with @xmath529 tables and @xmath530 chairs per table . guests who like each other should sit at the same table ; guests who dislike each other should not sit at the same table . the aggregate used in the problem is @xmath326 . we use the same setting to the problem instances as in @xcite . the results are shown in table [ seating ] . `` tables '' and `` chairs '' are the number of tables and the number of chairs at each table , respectively . the instance size is the number of atom occurrences in a ground program . we report the result of the average over one hundred randomly generated instances for each problem size . the experiments show that , by encoding logic programs with aggregates as weight constraint programs , alparse solves the problem efficiently . for large instances , the running time of alparse is about one order of magnitude lower than that of dlv and the sizes of the instances are also smaller than those in the language of dlv . we use the benchmarks reported in an asp solver competition and run all instances for each benchmark . in the experiments , we set the cutoff time to 600 seconds . the instances that are solved in the cutoff time are called `` solvable '' , otherwise `` unsolvable '' . table [ alparse - summary ] is a summary of the results . in the table , the time is the average running time in seconds for the solvable instances . it can be seen that alparse constantly outperforms dlv by several orders of magnitude , except for the benchmark of towers of hanoi . the system clasp has progressed to support aggregates @xmath325 , @xmath315 and @xmath294 . the aggregates used in the benchmarks are @xmath325 except for towers of hanoi where the aggregate @xmath294 is used . the aggregate @xmath325 is essentially the same as weight constraints . we compare the clasp programs with the aggregate @xmath294 and the corresponding translated weight constraint programs ( note that , the answer sets of this aggregate program correspond to those of the corresponding weight constraint program ) . the performances of clasp on these two kinds of programs are similar . as we have mentioned , the transformation approach indicates that it is important to focus on an efficient implementation of aggregate @xmath325 rather than on the implementation of other aggregates one by one , since they can be encoded by @xmath325 . can be translated to @xmath325 , using a logarithm transformation , thanks to tomi janhunen for the comments during the presentation of @xcite . ] .benchmarks used by smodels@xmath525 [ cols="<,>,>,>",options="header " , ] we have shown that for a large class of programs the stable model semantics coincides with the answer set semantics based on conditional satisfaction . in general , answer sets admitted by the latter are all stable models . when a stable model is not an answer set , it may be circularly justified . we have proposed a transformation , by which a weight constraint program can be translated to strong satisfiable program , such that all stable models are answer sets and thus well - supported models . we have also given another transformation from weight constraint programs to logic programs with nested expressions which preserves the answer set semantics . in conjunction with the one given in @xcite , their difference reveals precisely the relation between stable models and answer sets . as an issue of methodology , we have shown that most standard aggregates can be encoded by weight constraints . therefore the asp systems that support weight constraints can be applied to efficiently compute the answer sets of logic programs with aggregates . the experimental results demonstrate the effectiveness of this approach . currently , alparse does not handle programs with aggregates like @xmath531 or @xmath532 , due to the fact that the complexity of such programs is higher than @xmath533 . what is the best way to include this practically requires further investigation .

weight constraint and aggregate programs are among the most widely used logic programs with constraints . in this paper , we relate the semantics of these two classes of programs , namely the stable model semantics for weight constraint programs and the answer set semantics based on conditional satisfaction for aggregate programs . both classes of programs are instances of logic programs with constraints , and in particular , the answer set semantics for aggregate programs can be applied to weight constraint programs . we show that the two semantics are closely related . first , we show that for a broad class of weight constraint programs , called _ strongly satisfiable programs _ , the two semantics coincide . when they disagree , a stable model admitted by the stable model semantics may be circularly justified . we show that the gap between the two semantics can be closed by transforming a weight constraint program to a strongly satisfiable one , so that no circular models may be generated under the current implementation of the stable model semantics . we further demonstrate the close relationship between the two semantics by formulating a transformation from weight constraint programs to logic programs with nested expressions which preserves the answer set semantics . our study on the semantics leads to an investigation of a methodological issue , namely the possibility of compact representation of aggregate programs by weight constraint programs . we show that almost all standard aggregates can be encoded by weight constraints compactly . this makes it possible to compute the answer sets of aggregate programs using the asp solvers for weight constraint programs . this approach is compared experimentally with the ones where aggregates are handled more explicitly , which show that the weight constraint encoding of aggregates enables a competitive approach to answer set computation for aggregate programs . [ firstpage ] stable model , weight constraint , aggregates , logic programs with constraints .

there are several unique features in baryonic @xmath0 decays . first , a peak near the threshold area of the dibaryon invariant mass spectrum has been observed in many baryonic @xmath0 decays . second , three - body decays usually have rates larger than their two - body counterparts ; that is , @xmath20 . this phenomenon can be understood in terms of the threshold effect , namely , the invariant mass of the dibaryon is preferred to be close to the threshold . the configuration of the two - body decay @xmath21 is not favorable since its invariant mass is @xmath22 . in @xmath23 decays , the effective mass of the baryon pair is reduced as the emitted meson can carry away much energy . the low mass threshold effect can be understood in terms of a simple short - distance picture @xcite . for singly charmful baryonic @xmath0 decays , experimentally we have @xmath24 @xcite and @xmath25 @xcite . therefore , we have a pattern like @xmath26 where @xmath27 denotes a charmed baryon . the experimental measurements for doubly charmful @xmath0 decays are summarized in table [ tab:2bodycharm ] . for @xmath28 decays , we extract their branching ratios using @xmath29 and @xmath30 , respectively , @xmath31 where the second errors originate from the uncertainties in @xmath32 ranging from 0.83% to 1.74% and @xmath33 from 1.2% to 10.1% @xcite . theoretically , it is expected that the charged and neutral @xmath0 decays to @xmath34 should have similar rates . experimentally , this feature should be tested by the forthcoming measurements . since @xmath35 @xcite , we have another pattern @xmath36 for two - body baryonic @xmath0 decays . .branching ratios of doubly charmful two - body ( in units of @xmath37 ) and three - body ( in units of @xmath38 ) baryonic @xmath0 decays . [ cols="<,<,<",options="header " , ] in this work we have studied the doubly charmful two - body and three - body baryonic @xmath0 decays @xmath39 and @xmath19 . for the former decay , our prediction for its branching ratio of order @xmath17 ( see table [ tabc ] ) is consistent with the extrapolation from @xmath40 and from @xmath41 provided that the dynamical suppression of @xmath42 relative to @xmath6 is taken into account . as pointed out before , a naive estimate of the branching ratio @xmath3 for the decay @xmath19 is too small by three to four orders of magnitude compared to experiment . previously , it has been shown that a large enhancement for the @xmath4 production can occur due to a charmonium - like resonance ( for example , the @xmath5 state discovered by belle ) with a mass near the @xmath6 threshold . motivated by the babar s observation of a resonance in the @xmath43 system with a mass of order 2930 mev , we have studied the contribution to @xmath44 from the intermediate state @xmath9 which is postulated to be a first positive - parity excited @xmath10-wave charmed baryon state . assuming that a soft @xmath11 quark pair is produced through the @xmath12 and @xmath13 meson exchanges in the configuration for @xmath45 , it is found that the branching ratio of @xmath46 is of order @xmath16 . this is in agreement with experiment for @xmath47 , but slightly smaller for @xmath18 . in conjunction with the previous analysis @xcite , we conclude that the enormously large rate of @xmath19 arises from the resonances @xmath9 and @xmath5 . we have also presented updated results for @xmath48 which are slightly smaller than our previous analysis @xcite but are consistent with experiment . this work is supported in part by the national science council of r.o.c . under grants nsc97 - 2112-m-001 - 004-my3 and nsc97 - 2112-m-033 - 002-my3 . it has been conjectured that @xmath9 is likely to be a first positive - parity excitation with @xmath49 @xcite . denoting the quantum numbers @xmath50 and @xmath51 as the eigenvalues of @xmath52 and @xmath53 , respectively , the @xmath54-orbital momentum @xmath50 describes relative orbital excitations of the two light quarks , and the @xmath55-orbital momentum @xmath51 describes orbital excitations of the center of the mass of the two light quarks relative to the heavy quark @xcite . the first positive - parity excitations are those states with @xmath56 . according to table iv of @xcite , possible antitriplet candidates for @xmath9 are @xmath57 and @xmath58 , where the quantum number in the subscript labels @xmath59 , the total angular momentum of the two light quarks . ( we use a tilde to denote states with antisymmetric orbital wave functions ( i.e. @xmath60 ) under the interchange of two light quarks . ) strong decays of these four states have been studied in @xcite using the @xmath61 model . it turns out that @xmath62 mev is too small and @xmath63 mev is too large compared to the experimental value of order 30 mev [ see eq . ( [ width ] ) ] , while @xmath64 does not decay into @xmath65 and @xmath43 . therefore , the favored candidate is @xmath58 which has @xmath66 and @xmath67 . we use the light - front approach to obtain the structure of the matrix element for the @xmath10-wave charmed baryon as shown in eq . ( [ bwf1 ] ) . in the light - front formalism , the charmed baryon bound state with the total momentum @xmath68 , spin @xmath69 and the angular momentum of the light quark pair @xmath70 with @xmath71 and @xmath72 can be written as ( see , for example @xcite ) @xmath73 where @xmath74 and @xmath75 are color and flavor indices , respectively , @xmath76 denote helicities , @xmath77 , @xmath78 and @xmath79 are the on - mass - shell light - front momenta , @xmath80 and @xmath81 the coefficients @xmath82 and @xmath83 are normalized color factor and flavor coefficient , respectively . in terms of the light - front relative momentum variables @xmath84 for @xmath85 defined by @xmath86 the momentum - space wave - function @xmath87 can be expressed as @xmath88 where @xmath89 describes the momentum distribution of the constituents in the bound state with the subsystem consisting of the particles 2 and 3 in the orbital angular momentum @xmath90 state , @xmath91 , and so on are clebsch - gordan coefficients and @xmath92 is the well normalized melosh transform matrix element . explicitly @xcite , _ i|r^_m ( x_i , k_i , m_i)|s_i & = & = - + & = & , with @xmath93 and @xmath94 dirac spinors in the light - front ( instant ) form , @xmath95 , a unit vector in the @xmath96-direction , and m_0 ^ 2&=&_i=1 ^ 3 , k_i=(,x_i m_0 , k_i)=(e_i - k_iz , e_i+k_iz , k_i ) , + m_0&=&e_1+e_2+e_3 , e_i = = + , k_iz=- . + note that @xmath97 and , consequently , the state @xmath98 transforms like @xmath99 under rotation , i.e. its transformation does not depend on its momentum . a crucial feature of the light - front formulation of a bound state , such as the one shown in eq . ( [ lfmbs ] ) , is the frame - independence of the light - front wave function @xcite . namely , the hadron can be boosted to any ( physical ) ( @xmath100 , @xmath101 ) without affecting the internal variables ( @xmath102 , @xmath103 ) of the wave function , which is certainly not the case in the instant - form formulation . in practice it is more convenient to use the covariant form for the melosh transform matrix element @xmath104 with @xmath105 and @xmath106 , \quad \vec \varepsilon_\bot ( \pm 1)=\mp(1,\pm i)/\sqrt{2 } , \nonumber\\ & & \varepsilon^\mu(\bar p,0)={1\over m_0}\left({-m_0 ^ 2+p_\bot^2\over p^+},p^+,p_\bot\right ) , \label{polcom}\end{aligned}\ ] ] for states with @xmath107 or @xmath108 , see @xcite for the derivation the above expressions . the following identities will be useful later : & & 11;m m|11;2 m 12;m m|12;1 n _ ( |p , m)_(|p , m)_(|p , m ) + & & = - + & & + ^*_(|p , m)_(|p , m)_(| p , n ) , + & & ^*_(|p , m)_(|p , m ) = -g_+ , where @xmath109 is used in the first identity , and @xmath110 is defined as @xmath111 $ ] for later purposes . under the constraint of @xmath112 , we have the expressions & & _ l_kl_k , l_kl_k(\{x},\{k _ } ) = ( ) ^3/2_l_k l_k(k,_k ) _ l_k l_k(k,_k ) , + & & _ 00(,)= ( , ) , _ 1m(,)=_m _ p ( , ) , [ eq : phi ] where @xmath113 , or explicitly @xmath114 , @xmath115 with @xmath116 in the rest frame of @xmath117 , are proportional to the spherical harmonics @xmath118 in the momentum space , and @xmath119 , @xmath120 are the distribution amplitudes of @xmath121-wave and @xmath68-wave states , respectively , and the factor @xmath122 in eq . ( [ eq : phi ] ) is a normalization factor . for a gaussian - like wave function , one has @xcite @xmath123 by using the above equations it is straightforward to obtain & & ( _ i_i|r_m^(x_i , k_i , m_i)|s_i ) j_l s_1 ; j_l s_1|j_l s_1;j j_z s_l l_l ; s_l l_l|s_l l_1;j_l j_l + & & l_k l_k ; l_k l_k|l_k l_k;l_l l_l s_2 s_3 ; s_2 s_3|s_2 s_3;s_l s_l + & & = |u(p_1,_1)u_*b_c*(|p , j_z ) |u(p_2,_2)(|p+m_0)_5 c | u^t(p_3 , _ 3 ) [ eq : s ] for @xmath124 and & & ( _ i_i|r_m^(x_i , k_i , m_i)|s_i ) j_l s_1 ; j_l s_1|j_l s_1;j j_z s_l l_l ; s_l l_l|s_l l_1;j_l j_l + & & l_k l_k ; l_k l_k|l_k l_k;l_l l_l s_2 s_3 ; s_2 s_3|s_2 s_3;s_l s_l + & & = |u(p_1,_1)_5_u_*b_c*(|p , j_z ) + & & |u(p_2,_2)(|p+m_0 ) \{(^+^ ) -(^- ) } c | u^t(p_3 , _ 3 ) [ eq : d ] for @xmath125 , @xmath126 . note that the factors of @xmath127 , and @xmath128 come from the wave function eq . ( [ eq : phi ] ) for the @xmath129 case . promoting @xmath130 and @xmath131 and taking hermitian conjugation ( to change the initial state to the final state ) we obtain the structure of the matrix element for the @xmath121-wave and @xmath10-wave charmed baryons as shown in eq . ( [ bwf1 ] ) . note that for simplicity we have taken @xmath132 in eq . ( [ bwf1 ] ) . s. a. dytman _ et al . _ [ cleo collaboration ] , phys . rev . d * 66 * , 091101 ( 2002 ) ; n. gabyshev _ et al . _ [ belle collaboration ] , phys . rev . d * 66 * , 091102 ( 2002 ) ; n. gabyshev _ et al . _ [ belle collaboration ] , phys . rev . lett . * 97 * , 242001 ( 2006 ) ; k. s. park _ et al . _ [ belle collaboration ] , phys . rev . d * 75 * , 011101 ( 2007 ) . krner , m. krmer , and d. pirjol , prog . * 33 * , 787 ( 1994 ) . h. y. cheng , c. k. chua and c. w. hwang , phys . d * 69 * , 074025 ( 2004 ) . h. y. cheng , c. k. chua and c. w. hwang , phys . rev . d * 70 * , 034007 ( 2004 ) . h. y. cheng , c. y. cheung , and c. w. hwang , phys . d * 55 * , 1559 ( 1997 ) .

we study the doubly charmful two - body and three - body baryonic @xmath0 decays @xmath1 and @xmath2 . as pointed out before , a naive estimate of the branching ratio @xmath3 for the latter decay is too small by three to four orders of magnitude compared to experiment . previously , it has been shown that a large enhancement for the @xmath4 production can occur due to a charmonium - like resonance ( e.g. @xmath5 discovered by belle ) with a mass near the @xmath6 threshold . motivated by the babar s observation of a resonance in the @xmath7 system with a mass of order 2930 mev , we study in this work the contribution to @xmath8 from the intermediate state @xmath9 which is postulated to be a first positive - parity excited @xmath10-wave charmed baryon state . assuming that a soft @xmath11 quark pair is produced through the @xmath12 and @xmath13 meson exchanges in the configuration for @xmath14 and @xmath6 , it is found that branching ratios of @xmath15 and @xmath1 are of order @xmath16 and @xmath17 , respectively , in agreement with experiment except that the prediction for the @xmath18 is slightly smaller . in conjunction with our previous analysis , we conclude that the enormously large rate of @xmath19 arises from the resonances @xmath9 and @xmath5 . = cmbx10 scaled 2june , 2009 1.5 cm

the search for novel quantum states in strongly correlated electronic systems exhibiting an exotic order is at the core of modern condensed matter physics . an example that attracts strong recent interest is the _ nematic phase_. analogous to the nematic state observed in liquid crystals , electronic correlation in a nematic phase develops a _ preferred orientation _ , which breaks rotational symmetry . however , time reversal symmetry is preserved unlike in a conventional magnetic order . the possibility of such nematic order has been discussed in several different systems such as frustrated magnets,@xcite iron pnictide superconductors,@xcite and heavy fermion materials.@xcite among them , one dimensional ( 1d ) frustrated quantum spin systems ( @xmath17 ) having a ferromagnetic nearest neighbor ( nn ) exchange interaction @xmath18 and an antiferromagnetic next nearest neighbor ( nnn ) interaction @xmath19 , @xmath20 in magnetic fields deserve particular attention since extensive theoretical efforts have been devoted to this model.@xcite such intrachain interactions together with ferromagnetic interchain coupling in the case of large spin values yield a rich magnetic phase diagram.@xcite in particulary , the case @xmath21 leads to a helical magnetically ordered structure . in principle , a true long range magnetically ordered structure of a 1d chain of magnetic moments is suppressed by spin fluctuations , and a magnetic state of such a system is characterized by short range correlations between the magnetic moments . but according to theory , a long range magnetic order can be stabilized due to interactions beyond eq . ( [ j1j2 ] ) , i.e. , interactions with an applied magnetic field , interactions with the crystalline environment , or interactions between magnetic chains . for instance a moderate static magnetic field @xmath1 stabilizes helical correlations in favour of a long range ordered state even in one dimension@xcite with the order parameter @xmath22 for @xmath17 and @xmath23 . in contrast to the helical magnetic structure the expectation values of the transverse spin components @xmath24 and @xmath25 are equal to zero . such magnetic state with this pseudovector order parameter was labeled the chiral phase in the literature . to our knowledge , this chiral phase has not been observed experimentally . one dimensional frustrated quantum spin systems with competing exchange interactions given by eq . ( [ j1j2 ] ) are intriguing also in high magnetic fields , near the saturation field @xmath3 . let us first consider the fully polarized ground state above the saturation field @xmath26 . in ordinary unfrustrated antiferromagnets , the lowest energy excitation is described by a single spin flip , or a magnon , which acquires kinetic energy due to exchange interactions , forming a dispersive magnon band . as the field is reduced , the excitation gap determined by the minimum energy of the magnon band vanishes at @xmath3 . below @xmath3 , the bose - einstein condensation ( bec ) of magnons occurs in the presence of finite three dimensional interchain coupling , resulting in an antiferromagnetic order in the plane perpendicular to the field with the order parameter @xmath27 ( ref . ) . in contrast , the lowest energy excitation in the frustrated chain of eq . ( [ j1j2 ] ) immediately above @xmath3 is a bound magnon pair , not a single spin - flip , stabilized by the ferromagnetic nn interaction for a wide range of the ratio @xmath28 ( refs . ) . below @xmath3 , these bound magnon pairs exhibit bec in the presence of special type interchain coupling and establish a long range ordered magnetic structure with the order parameter given by @xmath29 ( refs . ) . this type of magnetic order in the recent literature is labeled as _ nematic_. again , the nematic order breaks the spin rotational symmetry . more precisely , the preferred orientation of the nearest neighbor spin correlation ( director ) rotates @xmath30 from one bond to the next . however , the transverse components @xmath24 and @xmath25 are zero and the longitudinal magnetization remains uniform , @xmath31 , where @xmath32 is the bulk magnetization and @xmath33 is the number of magnetic ions of the sample . as the field is further decreased , theories predict that the nematic order is replaced by a spin density wave ( sdw ) state , where the moments are collinear with the external field and their magnitudes are modulated with a generally incommensurate periodicity.@xcite this sdw state can be regarded as a spatial order of the bound magnon pairs stabilized by their mutual repulsive interaction , which becomes dominant over the gain in the kinetic energy of the bec phase as the density of bound magnon pairs increases . measured in pulsed magnetic fields @xmath1 oriented @xmath34 at @xmath35 k in licuvo@xmath0 . inset : magnification of @xmath11 for fields around the saturation field @xmath3 . data of three different single crystals s1 , s2 and s0@xmath36 are shown , where the data of s0@xmath36 are taken from ref . . the dashed lines are parallel to each other and guide the eye . the @xmath11 data for the samples s1 and s2 are more noisy than for s0@xmath36 , because these samples are significantly smaller.,width=85 ] a promising candidate compound where these theoretical predictions are expected to be tested experimentally is the antiferromagnet licuvo@xmath0 , which contains frustrated chains of cu@xmath9 ( @xmath17 ) along the crystallographic @xmath37direction.@xcite let us first summarize the results of the previous studies on the magnetic phases of licuvo@xmath0 for fields lower than @xmath38 t ( refs . ) . at zero field , an incommensurate planar spiral structure is realized below @xmath39 k with the moments lying in the @xmath40plane.@xcite a spin flip transition occurs at @xmath41 t , flipping the moments into the plane perpendicular to the field.@xcite in higher fields @xmath42 ( @xmath43 t ) a collinear spin modulated structure is realized,@xcite consistent with the theoretical prediction.@xcite remarkably , the relation between the magnetization and the wave vector of the spin modulated structure measured by neutron diffraction experiments in fields up to 15 t demonstrates that it is the bound magnon pairs with @xmath44 that form a periodic structure.@xcite also the temperature dependence of the nuclear spin lattice relaxation rate @xmath45 at the v sites indicates the development of an energy gap in the transverse spin excitation spectrum above @xmath46,@xcite which corresponds to the binding energy of magnon pairs , as theoretically predicted.@xcite the observation of the spin modulated phase and the experimental proof for the bound magnon pairs provide strong support that licuvo@xmath0 is indeed well described by the model of eq . ( [ j1j2 ] ) . by further increase of the applied magnetic field @xmath1 , it is expected that the spin nematic phase develops before the magnetization saturates at @xmath3 . in recent experiments the field dependent magnetization curve @xmath11 of licuvo@xmath0 exhibited anomalies slightly below @xmath47 t , indicating a new phase.@xcite figure 1 shows @xmath11 for three different samples measured in pulsed magnetic fields @xmath1 applied along the @xmath2direction at @xmath35 k. in addition to the anomalies at @xmath48 and @xmath3 already mentioned , all samples show another anomaly at @xmath49 t , about 5 t below the saturation . the sharp increase of the magnetization towards higher field stops at @xmath50 . the magnetization then increases linearly in the field range @xmath51 with the slope @xmath52 . this magnetization behavior near saturation including the slope of @xmath11 for @xmath51 is very similar for all sample . in this paper we discuss the nmr spectra of @xmath7v nuclei and the nuclear spin lattice relaxation rate @xmath45 of @xmath53li nuclei for applied magnetic fields @xmath1 along the @xmath2direction ( @xmath34 ) near the saturation field @xmath54 t. the nmr spectra represent distributions of the internal magnetic fields which are produced by the surrounding magnetic cu@xmath9 moments . the internal field has a large distribution in the spin modulated phase leading to broad nmr spectra . in the nematic phase , in contrast , moments are uniform and we expect a sharp single , solitary nmr line similar to the saturated phase . however , unlike in the saturated phase , the internal field determined from the peak position of the spectrum should change with the magnetic field @xmath1 as it is the case for the bulk magnetization @xmath11 in the nematic phase . such expectation may not be realized if the sample contains defects , which generate inhomogeneous distribution of the magnetic moments . since the nmr spectra reflect the histogram of magnetic moments while @xmath11 gives average magnetization , they may behave differently as we discuss below . the observation of the phase transition from the spin - modulated phase to the magnetically saturated phase from a local point of view is the major subject of the present work . in addition to @xmath11 measurements , which give bulk properties , we employed nmr of @xmath7v ( @xmath55 , @xmath56 mhz / t ) nuclei probing the local magnetic properties of the cu@xmath9 moments in licuvo@xmath0 . three single crystalline samples were used in the nmr measurements : sample s0 , s1 , and s2 . sample s0 was used in the previous nmr experiments in refs . and obtained from the same batch as s0@xmath36 whose @xmath11 curve is shown in fig . [ figure-1 ] . samples s1 and s2 are from a new batch whose @xmath11 curves are also shown in fig . [ figure-1 ] . echo techniques were utilized in the hybrid 45 t magnet of the dc field facility at the national high magnetic field laboratory ( nhmfl ) , tallahassee , florida . the absolute values of the applied magnetic fields @xmath1 were calibrated by nmr using aluminum powder . all spectra of sample s0 were collected with the pulse sequence @xmath57 ( @xmath58 ) by sweeping the applied magnetic field @xmath1 at constant frequencies . spectra of the single crystalline samples s1 and s2 were recorded by summing the fourier transform of the spin - echo signal with the pulse sequence @xmath59 ( @xmath60 ) obtained for equally spaced excitation frequencies at a fixed magnetic field . temperatures down to @xmath61 mk were achieved within a @xmath62he cryostat . v nmr spectra obtained from crystal s2 at @xmath61 mk within the field range @xmath63 t for @xmath34 . the spin - echo amplitudes are normalized by the peak intensity . the spectral shape shows a crossover at @xmath49 t from the double - horn pattern at lower fields to the single - peak pattern at higher fields . the dashed arrow in the inset denotes the expected line shift which is estimated from the bulk magnetization @xmath11.,width=85 ] v nmr spectra in licuvo@xmath0 are plotted against applied field along the @xmath2direction obtained from samples s0 ( squares ) , s1 ( triangles ) , and s2 ( bullets ) , respectively ( left axis ) . the solid line shows the bulk magnetization @xmath11 vs. the applied magnetic field @xmath1 obtained from sample s0@xmath36 ( right axis ) . inset : @xmath7v nmr spectra of sample s2.,width=85 ] figure [ figure-2 ] shows the nmr spectra of sample s2 obtained at @xmath61 mk within the range @xmath63 t of the applied magnetic field @xmath1 for @xmath34 . note that @xmath64 represents the effective magnetic field sensed by nuclei being observed at the resonance frequency @xmath65 and is the sum of the applied field plus the internal local field due to cu@xmath9 moments , i.e. , @xmath66 . therefore , the abscissa axis of the spectra in fig . [ figure-2 ] @xmath67 is equal to @xmath68 at the probing @xmath7v nuclei . the double - horn shaped pattern for lower fields @xmath1 at the bottom of fig . [ figure-2 ] resemble our previous results,@xcite where we established the spin - modulated magnetic structure in licuvo@xmath0 . the asymmetry of the double horn shaped pattern for @xmath69 t turned out to be due to different spin spin relaxation times @xmath70 at different spectral positions.@xcite around the value of the applied magnetic field @xmath71 t the double horn shape starts to disappear in favor of a single peak spectrum evolving a more symmetric line shape towards higher magnetic fields @xmath1 . for applied magnetic fields @xmath1 higher than @xmath10 t , the internal field occurs to stay almost constant without any shift in the peak position of the spectral line . this observation is in contrast to the @xmath11 data which further increase linearly for the same field values up to the saturation field @xmath3 ( cf . the inset in fig . [ figure-1 ] ) . this conflict is illustrated in the inset of fig . [ figure-2 ] , where the almost unshifted nmr spectral lines are plotted together with the internal field fields calculated from the @xmath11 data ( dashed arrow ) . the shaded sector in the inset of fig . [ figure-2 ] shows the estimated error made for the slope of the @xmath11 dependence obtained from pulse field experiments on @xmath11 ( cf . [ figure-1 ] ) . such unexpected behavior was observed for all three studied samples . the field dependences of the peak positions of the @xmath7v nmr spectra are shown in fig . [ figure-3 ] together with the @xmath72 data . the internal field at the vanadium nuclei for all three samples shows almost no change with field for @xmath73 t in contrast to the full magnetic moment of the sample @xmath11 measured in the same field range . this discrepancy between nmr line shift and bulk magnetization @xmath11 demonstrates that the magnetization is not uniform , probably due to the presence of defects . the constant internal field at the peak of the intense nmr line indicates that the majority part of the nuclei far from defects is embedded in an electronic surrounding with saturated magnetic moments , while the bulk magnetization is affected by defects . this point is discussed in more detail below . therefore , the nmr result is seemingly ineffective in detecting the existence of a spin nematic phase expected within the field range @xmath74 previously . however , close inspection of the data in fig . [ figure-3 ] reveals that within a limited narrow field range @xmath75 t ( indicated by the shaded area in fig . [ figure-3 ] ) the nmr spectral pattern is characterized by a nearly symmetric , solitary , single peak whose resonance position shifts strongly depending on the applied magnetic field @xmath1 ( cf . the inset of fig . [ figure-3 ] ) . this observation meets the theoretically predicted behavior of the spin nematic phase . indeed , the behavior of the nmr internal field within the restricted field range mentioned above is consistent with the theoretical prediction of a very steep slope of the magnetization curve in the nematic phase immediately below saturation.@xcite at @xmath14li nuclei in arrhenius representation measured for different applied magnetic fields @xmath76 t and 45.02 t. inset : the slopes of the arrhenius representations , i.e. activation energies @xmath77 , plotted vs. the applied magnetic field @xmath1 . the solid line indicates a linear fit ( see text ) . , width=85 ] at @xmath14li nuclei for different temperatures.,width=85 ] after the description of static magnetic properties probed by @xmath7v nmr mentioned above , we now turn to dynamic properties and the results of the nuclear spin lattice relaxation rate @xmath45 of @xmath14li nuclei ( @xmath78=3/2 , @xmath79 mhz / t ) . the sample used for the @xmath45 measurements is none of the three samples used for the @xmath7v nmr , but is from the same batch as the sample used in one of the previous nmr experiments described in ref . . figure [ figure-4 ] shows the temperature dependence of @xmath45 at two different applied magnetic fields above @xmath80 t. the arrhenius representation of the data exhibits a thermally activated behavior at low temperatures @xmath81 the activation energies are obtained as @xmath82 k at 45.02 t and @xmath83 k at 43.27 t ( see the inset of fig . [ figure-4 ] ) . assuming that the activation gap depends linearly on the field @xmath1 , @xmath84 we obtain @xmath85 k / t . in ordinary antiferromagnets , where the lowest energy excitations in the saturated state are single spin flip processes involving the change of magnetization by @xmath86 , the excitation gap should be given by @xmath87 with @xmath88 k / t ( @xmath89 = 2.31 , ref . ) . remarkably , the value of @xmath90 is much larger than @xmath91 but close to @xmath92 k / t . this indicates that the lowest energy excitations are bound magnon pairs with @xmath93 . the assumption of the linear field dependence of the energy gap is indeed justified by the exponential field dependence of @xmath45 above 41 t @xmath94 as shown in figure [ figure-5 ] . from the plots for two different temperature values we obtain @xmath95 t@xmath96 for @xmath97 k and @xmath98 t@xmath96 for @xmath99 k. by relating eq . ( [ exp_field ] ) to eqs . ( [ activation ] ) and ( [ gap ] ) , we obtain @xmath100 k / t ( 3.2 k / t ) for @xmath97 k ( @xmath99 k ) . these values are again close to @xmath101 , consistent with the picture that the lowest excitation is a two magnon bound state . from the plot in the inset of fig . [ figure-4 ] , we expect the excitation gap to become zero at applied magnetic fields @xmath102 t. this value is different from @xmath103 t determined from the magnetization data but agrees with the field value above which the @xmath7v nmr spectral peak stays constant . this fact provides further support for our previous conclusion that a majority of the magnetic copper ions in licuvo@xmath0 are already saturated _ below _ @xmath103 t. v spin - echo amplitude maxima of the double - horn pattern shown in fig . [ figure-2 ] ( red symbols are data taken from ref . ) . the other symbols denote the data of samples s0 ( black ) , s1 ( blue ) , and s2 ( green ) , respectively . the bullets denote the arithmetic average of high - field ( triangles ) and low - field ( squares ) maxima . the dashed lines guide the eye . right axis : the black solid line is the magnetization measurement @xmath11 of sample s0@xmath36 taken from ref . .,width=85 ] we first discuss the @xmath7v nmr spectra in the field range @xmath104 , where the magnetically ordered structure is recognized as the spin modulated structure . the characteristic double horn shape ( cf . fig . [ figure-2 ] ) is an unambiguous fingerprint of an incommensurate magnetic structure only observed within this field range . for lower applied magnetic fields @xmath105 or higher fields @xmath106 the @xmath7v nmr spectra consist of a single peak . the field dependence of the internal fields at the spectral maxima of the low field and high field horn in the case of the double - horn pattern , as well as the internal field of the singly peaked spectra are plotted in fig . [ figure-6 ] ( left axis ) . the arithmetically averaged field values @xmath107 ( solid bullets ) of the high field ( triangles ) and low field ( squares ) maxima , together with the single peak maxima tightly follow the field dependence of the magnetization data @xmath11 ( right axis in fig . [ figure-6 ] ) . -axis for different field ranges , illustrating cu@xmath9 ions ( large open circles ) , v@xmath108 ions ( small blue circles ) , and magnetic moments of cu@xmath9 ( arrows ) . panel a : the sinusoidal modulation below @xmath109 reconstructed from the values of @xmath110 and @xmath111 obtained by fitting the @xmath7v nmr spectrum at @xmath112 t , just below @xmath109 . panel b : illustration of intrachain exchange bonds with a defect replacing a magnetic ion by a nonmagnetic one ( gray bullet ) . panels c and d : models of the magnetic structure with a defect for fields @xmath113 and @xmath114 , respectively . the dashed circles around the non magnetic defect mark the region where the internal field of @xmath7v nuclei is smaller than the majority because the neighboring cu moments are unsaturated.,width=80 ] v nmr spectra of sample s0 obtained at @xmath115 mk for different applied magnetic fields in the range @xmath116 t. inset : data for sample s2 are plotted additionally ( solid line ) . the base lines of the different spectra in the inset are shifted along the @xmath117axis for better visibility.,width=80 ] v nmr spectra at @xmath115 mk for samples s0 and s2 measured at 483.3 and 475.4 mhz , respectively . solid lines show the fitting results of the main peak of the spectral pattern with a lorentzian line ( a1 ) . inset : relative integral intensity of the residual contribution a2 of the nmr spectra pattern which does not belong to the main line a1 . data from different samples s0 ( solid squares ) , s1 ( open triangle ) , and s2 ( open circles ) are shown.,width=80 ] we return to a closer inspection of the discrepancy between the internal field at the majority of vanadium nuclei obtained from nmr experiments and the bulk magnetization @xmath11 observed within the range of applied magnetic fields @xmath74 . the natural explanation for this phenomenon is the non uniform magnetization of the sample , i.e. , the magnetization of the majority part of the sample is saturated already at @xmath80 t , whereas there are local regions within the sample whose magnetization is saturated only at @xmath3 . in such a case , the intense line of the nmr spectra in the high field range is ascribed to the nuclei surrounded by magnetic moments with the saturated magnetization , whereas the experimentally observed growth of the bulk magnetization @xmath11 of all the samples can be attributed to the non saturated region near defects . the expected magnetic structures for the field ranges @xmath118 , @xmath74 , and @xmath114 are illustrated in figs . [ figure-7 ] a , c , and d , respectively . the structures in panel a ( below @xmath109 ) and c ( above @xmath119 ) both exhibit longitudinal modulation of the magnetization . in case of the spin modulated phase ( a ) this modulation has a regular sinusoidal shape , whereas for the high field phase ( c ) the modulation is not periodic but accidental . we expect that the latter structure is associated with crystallographic defects . although the nature of the defects is not precisely known , a composition study of licuvo@xmath0 indicates the presence of li defects of a few percents even in carefully prepared crystals.@xcite it was argued that a li deficiency produces a hole doped into the oxygen sites , which in turn will form a zhang rice singlet with a cu spin.@xcite such a singlet should be equivalent to a non magnetic defect replacing a cu spin and provokes an unusual magnetic state in its vicinity with deleted two nearest ferromagnetic exchange bonds , but with a conserved antiferromagnetic bond between the two parts of the interrupted chain ( see panel b of fig . [ figure-7 ] ) . a theoretical study shows that the saturation field of the magnetic neighborhood of such crystallographic defects is expected at distinctly higher fields compared to the saturation field of undisturbed magnetic chains.@xcite in the following we describe the fingerprints of such defect structure found in our experiments and evaluate the concentration of those defects . figure [ figure-8 ] is a replot of the field swept @xmath7v nmr spectra for applied magnetic fields @xmath1 between @xmath120 t at 380 mk . here , the ordinate axis is adjusted for each spectrum in such a way that the integrated intensity of the entire line is the same for all the spectra . this representation allows for an estimation of the defect concentration . the nmr spectra in this field range consist of an intense symmetric line and a broad plateau like part . the latter component is closer to the diamagnetic reference field at @xmath121 , and hence has smaller internal fields . the intense symmetric line can be fit by a lorentzian denoted a1 in fig . [ figure-9 ] , allowing us to separate the plateau like part marked as a2 . the inset of fig . [ figure-9 ] shows the field dependence of the relative weight of the plateau like part a2/(a1+a2 ) . for sample s0 this relative intensity of the plateau like part decreases monotonously within the entire field range @xmath122 , whereas for the samples s1 and s2 it was found to be smaller and to exhibit a weaker decrease with increasing field @xmath1 . the relative intensity a2/(a1+a2 ) at fields in the vicinity of @xmath3 is approximately 0.1 . from this value we can evaluate the defect concentration @xmath123 : the probability @xmath124 that all four magnetic sites surrounding the @xmath7v nucleus in its nearest proximity are occupied by cu@xmath9 ions is equal to @xmath125 . here we suppose that defects are distributed randomly . on the other hand , this value can be evaluated as @xmath126 . from this equation we obtain the concentration @xmath127 of non magnetic defects in reasonable agreement with the detailed investigation of reference . in conclusion , we investigated the magnetic structure of the cu@xmath9 moments in the quantum spin chains of licuvo@xmath0 in applied magnetic fields @xmath1 up to the saturation field @xmath128 t. the double horn shape of the @xmath7v nmr spectra in the incommensurate sdw phase changes to a single lorentzian line around @xmath129 t. although the magnetization curve @xmath11 shows a linear increase of magnetization in the field range @xmath74,@xcite the internal field corresponding to the peak of the nmr spectra stays constant for @xmath10 t , indicating that the moments surrounding the majority of vanadium nuclei are saturated in this field range . the results of the nuclear spin lattice relaxation rate of @xmath53li nuclei also show an energy gap expected for bound magnon pairs above the saturation at @xmath130 t. from these results , we conclude that the theoretically predicted nematic ordered phase can be realized only in the narrow field range @xmath131 t if it exists in licuvo@xmath0 . we have attributed the discrepancy between our nmr and magnetization data to effects caused by defects . from the careful investigation of a small pedestal like contribution in the @xmath7v nmr spectra , we were able to quantify the number of magnetic defects . as these defects locally yield much higher saturation fields @xmath3 we reconciled our nmr results with recent magnetization @xmath11 experiments.@xcite whether or not elimination of the defects stabilizes the spin nematic phase is an interesting future issue . this work is supported by the german research society ( dfg ) within the transregional collaborative research center trr 80 ( augsburg , munich ) , by the grants 13 - 02 - 00637 of the russian foundation for basic research , program of russian scientific schools , by a grant - in - aid for scientific research 25287083 from the japan society of promotion of science , and by the global ceo program ( core research and engineering of advanced materials - interdisciplinary education center for materials science)(no . g10 ) from mext , japan . work at the national high magnetic field laboratory is supported by the nsf cooperative agreement no . dmr0654118 , and by the state of florida . we thank h. tsunetsugu , n. shannon , a. smerald , m. sato , t. momoi , t. hikihara , h. t. ueda , a. matsuo , a. smirnov , s. sosin , o. starykh , and m. zhitomirsky for stimulating discussions . m. enderle , c. mukherjee , b. fk , r.k . kremer , j .- broto , h. rosner , s .- l . drechsler , j. richter , j. malek , a. prokofiev , w. assmus , s. pujol , j .- l . raggazzoni , h. rakoto , m. rheinstdter , and h.m . rnnow , europhys . lett . * 70 * , 237 ( 2005 ) .

we have performed nmr experiments on the quasi one dimensional frustrated spin-1/2 system licuvo@xmath0 in magnetic fields @xmath1 applied along the @xmath2axis up to field values near the saturation field @xmath3 . for the field range @xmath4 ( @xmath5 t and @xmath6 $ ] t ) the @xmath7v nmr spectra at @xmath8 = 380 mk exhibit a characteristic double horn pattern , as expected for a spin modulated phase in which the magnetic moments of cu@xmath9 ions are aligned parallel to the applied field @xmath1 and their magnitudes change sinusoidally along the magnetic chains . for higher fields , the @xmath7v nmr spectral shape changes from the double horn pattern into a single lorentzian line . for this lorentzian line , the internal field at the @xmath7v nuclei stays constant for @xmath10 t , indicating that the majority of magnetic moments in licuvo@xmath0 are already saturated in this field range . this result is inconsistent with the previously observed linear field dependence of the magnetization @xmath11 for @xmath12 with @xmath13 t [ l. e. svistov _ et al_. , jetp letters * 93 * , 21 ( 2011 ) ] . we argue that the discrepancy is due to non - magnetic defects in the samples . the results of the spin lattice relaxation rate of @xmath14li nuclei indicate an energy gap which grows with field twice as fast as the zeeman energy of a single spin , therefore , suggesting that the two magnon bound state is the lowest energy excitation . the energy gap tends to close at @xmath15 t. our results suggest that the theoretically predicted spin nematic phase , if it exists in licuvo@xmath0 , can be established only within the narrow field range @xmath16 t .

in cartography and data visualization , one often has to place similar copies of a symbol , typically an opaque disk , on a map or a figure at given locations @xcite , @xcite . the size of the symbol is sometimes proportional to the quantitative data associated with the location . on a cluttered map , it is difficult to identify the symbols . therefore , it has been investigated in several studies how to minimize the amount of overlap @xcite , @xcite . left : a collection of unit disks in the plane . right : a stacking order for them . ] in the present note , we follow the approach of cabello , haverkort , van kreveld , and speckmann @xcite . we assume that the symbols used are opaque circular disks of the same size . given a collection @xmath6 of @xmath0 distinct _ unit _ disks in the @xmath7-plane , a _ stacking order _ is a one - to - one assignment @xmath8 . we consider the integer @xmath9 to be the @xmath10-coordinate of the disk @xmath11 . the _ map _ corresponding to this stacking order is the 2-dimensional view of this arrangement from the point at negative infinity of the @xmath10-axis ( for notational convenience , we look at the arrangement from below rather than from above . ) in particular , for the lowest disk @xmath12 , we have @xmath13 , and this disk , including its full perimeter , is visible from below . the total length of the boundary pieces of the disks visible from below is the _ visible perimeter _ of @xmath6 with respect to the stacking order @xmath14 , denoted by @xmath15 . we are interested in finding a stacking order for which the visible perimeter of @xmath6 is as large as possible . see figure [ fig_stacking_order ] . there are other situations in which this setting is relevant . sometimes the vertices of a graph are not represented as points but as circles of a given radius . it may happen that some vertices overlap in the visualization ( especially if they have further constraints on their geometric position ) , and then it becomes important to choose a convenient stacking order that maximizes the visible perimeter . given an integer @xmath0 , we define @xmath16 where the maximum is taken over all stacking orders @xmath14 . we would like to describe the asymptotic behavior of @xmath17 , as @xmath0 tends to infinity . cabello _ et al . _ have already noted that @xmath18 ; in other words , every set @xmath6 of @xmath0 disks of unit radii admits a stacking order with respect to which its visible perimeter is @xmath19 . indeed , by a well - known result or erds and szekeres @xcite , we can select a sequence of @xmath20 disks @xmath21 such that their centers form a monotone sequence . more precisely , letting @xmath22 and @xmath23 denote the coordinates of the center of @xmath24 , we have @xmath25 and either @xmath26 or @xmath27 . then , in any stacking order @xmath14 such that @xmath28 for every @xmath29 , a full quarter of the perimeter of each @xmath30 is visible from below . therefore , the visible perimeter of @xmath6 with respect to @xmath14 satisfies @xmath31 at the problem session of _ eurocg11 _ ( morschach , switzerland ) , cabello , haverkort , van kreveld , and speckmann asked whether @xmath32 ; in other words , does there exist a positive constant @xmath5 such that every set of @xmath0 unit disks in the plane admits a stacking order , with respect to which its visible perimeter is at least @xmath33 ? we answer this question in the negative ; _ cf . _ theorems [ theorem2 ] and [ theorem5 ] below . given a set of points @xmath34 in the plane , let @xmath35 denote the collection of disks of radius @xmath36 centered at the elements of @xmath34 . for any positive real @xmath37 , let @xmath38 stand for a similar copy of @xmath34 , scaled by a factor of @xmath37 . for a stacking order @xmath14 of @xmath39 we will study the quantity @xmath40 . ( note the slight abuse of notation : we denote the stacking order of @xmath39 and the corresponding stacking order of @xmath41 by the same symbol @xmath14 . the two orders are also identified in lemmas [ lemma1 ] and [ lemma2.1 ] and in theorems [ theorem2 ] , [ theorem3 ] , and [ theorem5 ] . ) it is not hard to verify that , as @xmath37 gets smaller , the function @xmath40 decreases . to see this , it is enough to observe , as was also done by cabello _ et al . _ ( unpublished ) , that as we contract the set of centers , the part of the boundary of each unit disk visible from below shrinks . as we will see in lemma [ lemma2.1 ] , the limit in the following lemma has a simple alternative geometric interpretation . [ lemma1 ] for every point set @xmath34 in the plane and for every stacking order @xmath14 of the collection of disks @xmath35 , we have @xmath42 as in @xcite , @xcite , and @xcite , we consider _ @xmath43-dense _ @xmath0-element point sets @xmath34 , _ i.e_. , point sets in which the ratio of the maximum distance between two points to the minimum distance satisfies @xmath44 ( the above ratio is sometimes called the _ spread _ of @xmath34 @xcite ; thus , we consider point sets with spread at most @xmath45 . ) [ theorem2 ] for any @xmath43-dense @xmath0-element point set @xmath34 in the plane and for any stacking order @xmath14 , we have @xmath46 where @xmath47 is a constant depending only on @xmath43 . the order of magnitude of the upper bound in theorem [ theorem2 ] can not be improved : [ theorem3 ] for every positive integer @xmath0 , there exists a @xmath48-dense @xmath0-element point set @xmath49 in the plane and a stacking order @xmath14 such that @xmath50 in the general case , where @xmath34 is an arbitrary @xmath0-element point set in the plane , we have been unable to improve on the easy lower bound @xmath51 sketched above . however , under special assumptions on @xmath34 , we can do better . [ theorem4 ] every @xmath43-dense @xmath0-element point set @xmath34 in the plane admits a stacking order @xmath14 with @xmath52 where @xmath53 depends only on @xmath43 . in particular , theorem [ theorem4 ] provides an @xmath2 lower bound for the visible perimeter of a collection of @xmath0 unit disks centered at the points of an @xmath3 uniform grid , under a suitable stacking order . if the side length of the grid is very small , this is better than the line - by - line `` lexicographic '' stacking order , for which the visible perimeter is only @xmath54 . it turns out that in this case there is no stacking order for which the order of the magnitude of the visible perimeter would exceed @xmath55 . [ theorem5 ] let @xmath0 be a perfect square and let @xmath56 denote an @xmath57 by @xmath57 uniform grid in the plane . for any stacking order @xmath14 , we have @xmath58 consequently , we have @xmath59 . lemma [ lemma1 ] implies that the worst collections of disks are those whose centers are very close to each other , so all disks have a point in common . this is , of course , not a realistic assumption in the labeling problem in cartography that has motivated our investigations . in practical applications , only a bounded number of unit disks share a point . for such a case , we have the following result : [ thm_bounded_overlap ] let @xmath6 be a collection of @xmath0 unit disks in which at most @xmath5 disks have a point in common . then there exists a stacking order @xmath14 for which @xmath60 where @xmath61 is given in ( [ eq_def_v ] ) . this bound is worst - case asymptotically tight . in section 2 , we establish theorems [ theorem2 ] and [ theorem3 ] . the proof of theorem [ theorem4 ] is presented in section 3 . in section 4 , we consider the square grid and present a much simpler proof of this special case of theorem [ theorem4 ] based on jarnik s theorem @xcite ; we then prove theorem [ theorem5 ] , which states that the bound of theorem [ theorem4 ] is tight in this case . in section 5 , we prove theorem [ thm_bounded_overlap ] . the last section contains concluding remarks and open problems . in this section , we prove theorems [ theorem2 ] and [ theorem3 ] . first , we express the limit of visible perimeters in a simpler form . given a set of points @xmath34 in the plane , let @xmath62 stand for its convex hull . let @xmath63 denote the unit disk centered at @xmath64 and let @xmath35 stand for the set @xmath65 . fix an orthogonal system of coordinates in the plane . for any point @xmath66 and for any @xmath67 , let @xmath68 denote the point with coordinates @xmath69 . [ lemma2.1 ] let @xmath70 be a set of points in the plane , let @xmath67 , and let @xmath14 be the stacking order of @xmath71 given by @xmath72 for @xmath73 . we have @xmath74 where @xmath75 , and for all other indices , @xmath76 if @xmath77 belongs to @xmath78 , and @xmath79 is equal to the external angle of the convex polygon @xmath80 at vertex @xmath77 , otherwise . if @xmath77 lies outside the convex hull of the preceding points , then @xmath79 is defined as the external angle of the polygon @xmath81 at vertex @xmath77 . ] see figure [ fig_xang ] . we prove that the contribution of @xmath82 to the visible perimeter tends to @xmath79 as @xmath83 for each @xmath84 . since @xmath85 is the lowest disk in @xmath71 , its whole boundary is visible from below . therefore , its contribution is @xmath86 . let @xmath87 . if @xmath77 belongs to the interior of @xmath88 , then there is a threshold @xmath89 such that @xmath90 for every @xmath91 . in this case , no portion of the boundary of @xmath92 is visible from below , provided that @xmath37 is sufficiently small . if @xmath77 lies on the boundary of @xmath80 , then it is in between some points @xmath93 and @xmath94 with @xmath95 and although @xmath96 will not be entirely covered by earlier disks for any @xmath67 , the part of its boundary outside @xmath97 tends to zero as @xmath98 . finally , if @xmath77 lies outside @xmath99 , then it is a vertex of @xmath100 . consider the external unit normal vectors to the two sides of @xmath101 that meet at @xmath102 ( or in case the convex hull is a single segment , the two unit normal vectors for this segment ) . drawing these vectors from @xmath102 , the arc on the boundary of @xmath103 between them is of length @xmath79 and it is not covered by @xmath104 . thus , it is visible from below , and , as @xmath105 , the total contribution of the remaining part of the boundary of @xmath92 to the visible perimeter tends to @xmath106 , concluding the proof . for any @xmath43-dense @xmath0-element point set @xmath34 in the plane and for any stacking order @xmath14 , we have @xmath46 where @xmath47 is a constant depending only on @xmath43 . consider a @xmath43-dense point set @xmath34 in the plane and let @xmath14 be a stacking order for @xmath39 . using lemma [ lemma2.1 ] , it is enough to prove @xmath107 for the angles @xmath79 defined in the lemma . as @xmath76 whenever @xmath77 is contained in @xmath99 , we can assume this is never the case . since the quantity @xmath108 is independent of scale , we can assume without loss of generality that the minimum distance between points is @xmath36 ; thus , the maximum distance ( diameter ) is at most @xmath109 . we write @xmath70 with @xmath110 . the triangle @xmath111 lies entirely outside the convex hull of @xmath112 . ] for every @xmath113 , let @xmath114 denote the perimeter of @xmath80 . we define the perimeter of a segment to be twice its length and the perimeter of a point to be @xmath106 . let @xmath115 , consider the two sides of the polygon @xmath116 meeting at @xmath77 , and denote by @xmath117 and @xmath118 the points on these sides at unit distance from @xmath77 . since no point of @xmath34 is closer to @xmath77 than @xmath36 , the triangle @xmath111 does not contain any element of @xmath119 . ( see figure [ fig_perimeter ] . ) hence , @xmath120 is contained in the convex region obtained from @xmath121 by cutting off the triangle @xmath111 . ( in the degenerate case when @xmath100 is a segment , we have @xmath122 , and the empty `` triangle '' becomes just a unit segment . ) this observation implies that the perimeter of @xmath120 satisfies @xmath123 here we used that the external angle of the triangle @xmath111 at vertex @xmath77 is @xmath79 . thus , we have @xmath124 for all @xmath87 . adding up these inequalities , we obtain @xmath125 since @xmath126 is at most @xmath127 times the diameter of @xmath34 , that is , @xmath128 , we have @xmath129 applying the relationship between the arithmetic and quadratic means , we can conclude that @xmath130 taking into account that @xmath75 , the theorem follows by lemma [ lemma2.1 ] . for every positive integer @xmath0 , there exists a @xmath48-dense @xmath0-element point set @xmath49 in the plane and a stacking order @xmath14 such that @xmath50 a dense point set that has a good stacking order . ] suppose for simplicity that @xmath131 for some integer @xmath132 . our point set @xmath49 consists of the points having polar coordinates @xmath133 for @xmath134 and @xmath135 . see figure [ fig_concentric ] . the smallest distance between two points in @xmath49 is @xmath36 , and the largest distance is less than @xmath136 ; thus , @xmath49 is @xmath48-dense , as required . our stacking order @xmath14 takes the points by increasing @xmath137 , and for each @xmath137 by increasing @xmath138 . we apply lemma [ lemma2.1 ] and calculate the sum of the external angles determined by @xmath14 . denote by @xmath139 the circle of radius @xmath140 centered at the origin . consider a point @xmath141 on @xmath139 . let @xmath142 be the ray leaving @xmath64 towards the right tangent to @xmath139 , and let @xmath143 be the ray leaving @xmath64 towards the left tangent to @xmath144 . let @xmath117 be the point of tangency between @xmath143 and @xmath144 . then all the points of @xmath49 that precede @xmath64 in the order @xmath14 lie below @xmath142 and @xmath143 . thus , the external angle @xmath145 contributed by @xmath64 is at least the supplement @xmath146 of the angle between @xmath142 and @xmath143 . we have @xmath147 . the theorem follows . we now turn to theorem [ theorem4 ] . every @xmath43-dense @xmath0-element point set @xmath34 in the plane admits a stacking order @xmath14 with @xmath52 where @xmath53 depends only on @xmath43 . throughout this section , let @xmath34 be a @xmath43-dense @xmath0-point set in the plane . we will define a stacking order @xmath14 for @xmath39 for which the external angles @xmath79 defined in lemma [ lemma2.1 ] satisfy @xmath148 , for some constant @xmath53 depending only on @xmath43 . then the theorem follows from lemma [ lemma2.1 ] . assume without loss of generality that the minimum distance in @xmath34 is @xmath36 . then , since @xmath34 is @xmath43-dense , there exists a disk of radius @xmath109 that contains all of @xmath34 . let @xmath12 be such a disk , and let @xmath149 be a circle of radius @xmath150 concentric with @xmath12 . left : partition of @xmath12 into annular sectors centered at a point @xmath151 . top right : the sector containing @xmath77 is contained in the rectangle @xmath152 centered at @xmath77 . bottom right : point @xmath77 contributes external angle at least @xmath153 . ] given a point @xmath151 , we define a family @xmath154 of annular sectors that disjointly cover the plane , as follows : for each positive integer @xmath140 , let @xmath155 be a circle centered at @xmath64 with radius @xmath156 ; then divide each annulus between two consecutive circles into sectors of angular length @xmath157 for a large enough constant @xmath158 ( as will be specified below ) . see figure [ fig_annular_sectors ] ( left ) . note that each annular sector that intersects @xmath12 has area @xmath159 ( since the radius of such a sector is @xmath160 ) . the number of annular sectors that intersect @xmath12 is @xmath161 . call a sector _ occupied _ if it contains at least one point of @xmath34 . [ lemma3.1 ] there exists a point @xmath151 for which @xmath162 sectors of @xmath163 are occupied . choose @xmath64 uniformly at random on @xmath149 and construct the sectors using @xmath64 and dividing the annuli into the correct - length sectors in an arbitrary way . for each point @xmath164 , define the random variable @xmath165 to be the number of points of @xmath34 contained in the sector of @xmath163 that contains @xmath77 . we claim that the expected value @xmath166 $ ] of @xmath165 satisfies @xmath167 \le k\ ] ] for some constant @xmath168 . indeed , let @xmath169 be a rectangle centered at @xmath77 , with dimensions @xmath170 , and with short sides parallel to the line @xmath171 , for an appropriate constant @xmath172 . if @xmath172 is large enough ( but constant with respect to @xmath0 ) , then @xmath152 completely contains the sector of @xmath163 that contains @xmath77 . see figure [ fig_annular_sectors ] ( top right ) . thus , it suffices to bound the expected number of points of @xmath34 in @xmath152 . note that , as @xmath64 rotates around @xmath149 , @xmath152 rotates around its center together with @xmath64 . partition the plane into annuli centered at @xmath77 by tracing circles around @xmath77 of radii @xmath173 . the annulus with inner radius @xmath137 and outer radius @xmath174 contains at most @xmath175 points of @xmath34 , for some constant @xmath176 . each such point has probability at most @xmath177 of falling in @xmath152 ( over the choice of @xmath64 ) , for another constant @xmath178 ; therefore , the expected contribution of this annulus to @xmath165 is at most @xmath179 . summing up for all annuli with inner radius @xmath180 , we obtain that @xmath166 \le k$ ] for some constant @xmath168 , as claimed . now , call point @xmath77 _ isolated _ if @xmath181 . by markov s inequality , each point @xmath77 has probability at least @xmath182 of being isolated . therefore , the expected number of isolated points is at least @xmath183 . there must exist a @xmath64 that achieves this expectation , and for it we obtain at least @xmath184 occupied sectors , proving the lemma . fix a point @xmath64 for which @xmath163 has @xmath162 occupied sectors . color the sectors with four colors , using colors @xmath36 and @xmath185 alternatingly on the odd - numbered annuli and colors @xmath186 and @xmath48 alternatingly on the even - numbered annuli . there must be a color for which @xmath162 sectors are occupied . consider only the occupied sectors with this color . let these sectors be @xmath187 , listed by increasing distance from @xmath64 , and for each fixed distance , in clockwise order around @xmath64 . select one point @xmath188 from each of these sectors . let the stacking order @xmath14 start with these points , that is , @xmath189 for @xmath190 . the order of the remaining points in @xmath34 is arbitrary . we claim that each selected point @xmath77 contributes an external angle of @xmath191 , which implies that @xmath192 , as desired . indeed , consider the @xmath140-th selected point @xmath77 . suppose without loss of generality that @xmath64 lies directly below @xmath77 . let @xmath193 and @xmath194 be the inner and outer circles bounding the annulus that contains @xmath77 . trace rays @xmath195 and @xmath196 from @xmath77 tangent to @xmath197 , touching @xmath197 at points @xmath198 and @xmath199 . see figure [ fig_annular_sectors ] ( bottom right ) . every point @xmath93 , @xmath200 , that is _ not _ contained in the same annulus as @xmath77 lies below these rays . moreover , the angle @xmath153 that these rays make with the horizontal is @xmath201 : consider , for example , the ray @xmath195 . the triangle @xmath202 is right - angled , with angle @xmath203 . we have @xmath204 and @xmath205 . it follows that @xmath206 , and so @xmath207 . now suppose that @xmath208 lies in the same annulus as @xmath77 . if the constant @xmath158 in the definition of @xmath146 is chosen large enough , then @xmath208 must have a smaller @xmath209-coordinate than @xmath77 . ( in the worst case , @xmath77 lies near the bottom - left corner of its sector and @xmath208 lies near the top - right corner of its sector . ) thus , @xmath77 contributes external angle @xmath210 , as claimed . in this section , we assume that @xmath0 is a square number and @xmath56 denotes an @xmath57 by @xmath57 integer grid . note that @xmath56 is a @xmath211-dense set consisting of @xmath0 points . as we mentioned in the introduction , in the special case where @xmath212 , theorem [ theorem4 ] has a simple proof . for @xmath213 , one can produce a stacking order with large visible perimeter using the following greedy algorithm ( which can also be applied to any other point set @xmath34 ) : set @xmath214 , and select a vertex of @xmath215 whose external angle is maximum . let this vertex be @xmath216 , the last element in the desired order @xmath217 . repeat the same step for the set @xmath218 , and continue in this fashion until the first element @xmath219 gets defined . by jarnik s theorem @xcite , every convex polygon has @xmath220 vertices in @xmath56 . therefore , at each step , the greedy algorithm selects a point @xmath77 that makes an external angle @xmath191 . hence , @xmath221 for the order @xmath217 . lemma [ lemma2.1 ] completes the proof . now we turn to theorem [ theorem5 ] . let @xmath0 be a perfect square and let @xmath56 denote an @xmath57 by @xmath57 uniform grid in the plane . for any stacking order @xmath14 , we have @xmath58 our proof is an improved version of the proof of theorem [ theorem2 ] . there we were concerned with how the _ perimeter _ of the convex hull grows as we add the points of our set one by one as prescribed by the stacking order . as is well known , the perimeter of a convex set in the plane is the integral of its width in all directions ( this is known as cauchy s theorem ; see _ e.g. _ @xcite , theorem 16.15 ) . the proof of theorem [ theorem5 ] is very similar , but we deal with the widths in different directions in a non - uniform way . the width in a direction close to the direction of a short grid vector is more important in the analysis than widths in other directions . let @xmath222 be an enumeration of the points of @xmath56 according to a given stacking order , and let @xmath79 denote the corresponding external angles , as defined in lemma [ lemma2.1 ] . according to the lemma , we need to prove that @xmath223 . let us partition this sum into several parts , and bound the contribution of each part separately . let @xmath224=\{1,\ldots , n\}$ ] . we start with the small angles . let @xmath225\mid\tau_i < n^{-1/3}\}.\ ] ] clearly , we have @xmath226 the triangle @xmath227 is the largest isosceles triangle at point @xmath77 that does not intersect the interior of @xmath144 . ] as in the proof of theorem [ theorem2 ] , let @xmath228 and denote the perimeter of @xmath139 by @xmath114 . since @xmath56 is an @xmath3 integer grid , we have @xmath229 . consider only those indices @xmath87 that do not belong to @xmath230 . for these indices , we have @xmath231 , so that @xmath77 must be a vertex of @xmath139 . for each such point @xmath77 , let @xmath232 denote the smallest number satisfying the following condition : the segment connecting the points @xmath233 and @xmath234 that lie on the boundary of @xmath139 at distance @xmath232 from @xmath77 , intersects @xmath144 . ( in the case where @xmath139 is a segment , we have @xmath235 . ) note that the segment @xmath236 contains a point @xmath93 with @xmath237 . see figure [ fig_v_beta ] . in the proof of theorem [ theorem2 ] , we argued that @xmath238 . now the same argument gives that @xmath239 . let @xmath240\setminus(i_0\cup\{1\})\mid c_i\tau_i > n^{-1/6}\}.\ ] ] for @xmath241 , we have @xmath242 . since @xmath114 is monotone in @xmath140 , we conclude that @xmath243 let @xmath244\setminus(\{1\}\cup i_0\cup i_1).\ ] ] to bound the angles @xmath79 for indices @xmath245 , we need a charging scheme and we need to consider the growth of the width of @xmath139 in some specific directions . the _ width _ of a planar set in a given direction is the diameter of the orthogonal projection of the set to a line in this direction . let us associate the directions in the plane with the points of the unit circle @xmath149 . we identify opposite points of this circle as the widths of the same set in opposite directions are the same . this makes the total length of @xmath149 become @xmath127 . we define a set of arcs along @xmath149 as follows . for any non - zero grid vector @xmath246 from the integer grid and for any integer @xmath247 , let @xmath248 denote the arc of length @xmath249 symmetric around the direction of @xmath246 . for any direction @xmath250 , let @xmath251 denote the width of @xmath139 in the direction _ orthogonal _ to @xmath146 ( _ i.e. _ , where the corresponding projection is parallel to @xmath146 ) . the perimeter @xmath114 is equal to the integral of @xmath251 along the circle @xmath149 ( note that after the identification of opposite points the length of @xmath149 became @xmath127 ) . we have @xmath252 , unless the direction @xmath146 is tangent to @xmath139 at the vertex @xmath77 . let @xmath253 denote the arc of directions where such a tangency occurs . clearly , the length of @xmath253 is @xmath79 , and for any arc @xmath254 that contains @xmath253 , we have @xmath255 for each index @xmath245 , choose a grid point @xmath93 on the segment @xmath256 . ( recall that the points @xmath233 and @xmath257 are at distance @xmath232 from @xmath77 , and that there is always a grid point between them . ) we _ charge _ the index @xmath140 to the pair @xmath258 , where @xmath259 is the grid vector pointing from @xmath93 to @xmath77 and @xmath260 is the largest integer such that @xmath261 contains @xmath253 . notice that @xmath262 . denote by @xmath263 the set of indices @xmath245 that are charged to the pair @xmath264 . note that @xmath253 is symmetric around the direction of the segment @xmath256 . for the angle @xmath153 between this direction and the direction of @xmath259 we have @xmath265 ( refer again to figure [ fig_v_beta ] ) . this implies @xmath266 , and hence @xmath267 . finally , we also have @xmath268 let @xmath269 . the integral @xmath270 is monotone in @xmath140 and grows by at least @xmath271 at every @xmath272 . we have @xmath273 and @xmath274 , so that the final integral satisfies @xmath275 . therefore , @xmath276 , which implies that @xmath277 . consider the set of all pairs @xmath264 such that there is an index @xmath245 charged to them . we have @xmath278 , @xmath279 and @xmath280 , which implies that @xmath281 . we proved that @xmath282 . on the other hand , we also have @xmath283 . thus , for any given grid vector @xmath246 , there are at most @xmath284 possible values of @xmath142 , where @xmath285 denotes the binary logarithm . hence , @xmath286 to evaluate this sum , we note that the number of grid vectors @xmath246 satisfying @xmath287 is @xmath288 . thus , @xmath289 in conclusion , we have @xmath290 completing the proof of the theorem . in this section , we prove theorem [ thm_bounded_overlap ] . let @xmath6 be a collection of @xmath0 unit disks in which at most @xmath5 disks have a point in common . then there exists a stacking order @xmath14 for which @xmath60 where @xmath61 is given in ( [ eq_def_v ] ) . this bound is worst - case asymptotically tight . note that lemma [ lemma2.1 ] is not relevant in this case , since we can not contract the set of centers of @xmath291 . partition the plane into an infinite grid of axis - parallel square cells of side - length @xmath48 , where the position of the grid is chosen uniformly at random . for each unit disk , the probability that it belongs entirely to a single cell is @xmath292 . thus , we can fix the grid in such a way that at least @xmath293 disks lie entirely in a cell . let @xmath294 be the number of disks entirely contained in cell @xmath140 . by area considerations , we have @xmath295 . for each cell @xmath140 , we independently select a stacking order that achieves visible perimeter at least @xmath296 ; then we place all the remaining disks behind them . thus , our stacking order achieves visible perimeter at least @xmath297 . for any @xmath0-element point set @xmath6 , we can take an @xmath298-element point set @xmath299 as the union of @xmath137 pairwise disjoint translates of @xmath6 . we clearly have @xmath300 . this implies that @xmath301 . let @xmath302 , and we have @xmath303 , thus @xmath304 since @xmath305 , the claimed bound follows . to show that this bound is worst - case asymptotically tight , take the union of @xmath306 worst - case sets of @xmath5 disks far from each other . * a. * the greedy algorithm described at the beginning of section 4 was first considered by cabello _ et al . _ ( unpublished ) in the context of maximizing the _ minimum _ visible perimeter of a single disk . they showed that the order @xmath217 is always optimal for this purpose . unfortunately , this stacking order is _ not _ always optimal with respect to the total visible perimeter . indeed , let @xmath0 be a perfect square and consider the set of points @xmath307 , where the polar coordinates of @xmath77 are @xmath308 with @xmath309 sufficiently small . this point set is obtained as the intersection of @xmath57 equally - spaced rays emanating from the origin , and @xmath57 `` rounds '' of a very tight logarithmic spiral centered at the origin . the greedy algorithm produces the stacking order indicated by the indices , so it takes the points of @xmath34 outwards along the spiral . the contribution @xmath79 is equal for every point @xmath77 with @xmath310 and tends to @xmath311 as @xmath312 goes to zero , making @xmath313 if @xmath312 is small enough . however , taking the points ray by ray in a cyclic order , going outwards along each ray , the contribution @xmath79 is a constant for the first half of the points , making @xmath314 . * b. * theorem 4 can be generalized to point sets satisfying weaker density conditions . indeed , let @xmath34 be a set of @xmath0 points in the plane with diameter @xmath12 and minimum distance @xmath315 . a randomized construction , similar to the one used in the proof theorem 4 , guarantees the existence of a stacking order @xmath14 such that @xmath316 . this beats the @xmath19 bound mentioned in the introduction as long as @xmath317 .

given a collection of @xmath0 opaque unit disks in the plane , we want to find a stacking order for them that maximizes their _ visible perimeter _ , the total length of all pieces of their boundaries visible from above . we prove that if the centers of the disks form a _ dense _ point set , _ i.e. _ , the ratio of their maximum to their minimum distance is @xmath1 , then there is a stacking order for which the visible perimeter is @xmath2 . we also show that this bound can not be improved in the case of a sufficiently small @xmath3 uniform grid . on the other hand , if the set of centers is dense and the maximum distance between them is small , then the visible perimeter is @xmath4 with respect to any stacking order . this latter bound can not be improved either . finally , we address the case where no more than @xmath5 disks can have a point in common . these results partially answer some questions of cabello , haverkort , van kreveld , and speckmann . keywords : visible perimeter , disk , unit disk , dense set .

the most massive stars end their lives as core - collapse supernovae ( sne ) . the broad category of core - collapse sne is divided spectroscopically based on the presence or absence of h and he , with the sequence from strong h , to he and weak h , to only he , to lacking both h and he being ( respectively ) types ii , iib , ib , and ic ( see * ? ? ? * for a review ) . these correspond to progenitor stars with progressively decreasing h and he envelopes , with the mass loss caused by either stellar winds or mass transfer to a companion star . type iin sne @xcite are objects with narrow h emission lines , which are the result of the sn ejecta interacting with a dense circumstellar medium ( csm ) @xcite . wolf - rayet ( wr ) stars are evolved , core he - burning stars near the end of their lives . they are thought to be the descendants of stars that begin their lives with initial masses @xmath1@xmath2 , which have already shed their h envelopes during a luminous blue variable ( lbv ) stage @xcite , exposing their he cores ( for a review , see * ? ? ? * ) . during this stage in stellar evolution , the mass loss from stellar winds is @xmath0 @xmath3 yr@xmath4 , about 100 times higher than that of other o and b stars @xcite . although the wr stage lasts only @xmath0 @xmath5 years @xcite , a star can lose several solar masses of material during this stage . these facts suggest that wr stars are possible progenitors of stripped - envelope ( types iib through ic ) sne . @xcite showed that sn 1999cq had intermediate - width ( fwhm @xmath6 ) he emission lines similar to the hydrogen lines of sne iin , but lacking h lines . they suggested that sn 1999cq was interacting with a csm of dense he , but little or no h. until now , only one other supernova , sn 2002ao @xcite , has been identified as having intermediate - width he emission lines @xcite . sn 2006jc was discovered in by amateur astronomers @xcite on 2006 oct . 9.75 ( ut dates are used throughout this paper ) . a non - detection was reported on sep . 22 , suggesting that the sn was discovered shortly after explosion . soon after detection , several groups obtained spectra of sn 2006jc and noted the presence of emission lines , but did not associate it with sn 1999cq @xcite . @xcite later reclassified sn 2006jc as type ia , before @xcite noted its similarity to sne 1999cq and 2002ao . in addition to optical photometry and spectroscopy presented in this letter , uv and x - ray data were obtained with the _ swift _ satellite @xcite , with a positive x - ray detection , probably indicating cs interaction . further x - ray observations with _ chandra _ @xcite confirm the _ swift _ x - ray results . detailed study of the well - observed sn 2006jc presents an opportunity to examine the progenitor of a rare but important core - collapse sn spanning the gap between sne iin and ib , sampling the environment of a wr star at the end of its stellar life , and helping to clarify what appears to be a new subclass of sne . here we present optical photometry and spectroscopy and discuss the implications of those data . we also present observations of sn 2002ao in the interest of completeness , but leave a detailed analysis of all the data for a future paper . once the unusual nature of sn 2006jc was noted , we began a monitoring campaign that consisted of photometry with the katzman automatic imaging telescope ( kait ; * ? ? ? * ) and optical spectroscopy with the kast spectrograph @xcite mounted on the lick observatory 3-m shane telescope , the lris spectrograph @xcite mounted on the keck i 10-m telescope , and the deimos spectrograph @xcite mounted on the keck ii 10-m telescope . the multi - band photometry began on oct . 11 while our first spectrum was obtained on oct . 24 . the photometry in this paper continues through dec . 16 while our last spectrum in this paper was obtained on nov . 24 . all spectral data were reduced using standard techniques ( e.g. , * ? ? ? using our own idl routines , we fit spectrophotometric standard star spectra to flux - calibrate our data and remove telluric lines @xcite . photometric data were obtained with kait and the 1-m nickel telescope at lick observatory . magnitudes were measured in the kron - johnson system using the point - spread function ( psf)-fitting photometry software @xcite in the iraf daophot package , as multi - color template images required for galaxy subtraction are not available and the sn is still visible in kait data . since sn 2006jc is bright and reasonably isolated from its faint host galaxy , galaxy subtraction is not necessary , and simple psf - fitting provides us a reasonable approach to reduce the photometry at this early time . the instrumental magnitudes for sn 2006jc derived this way are calibrated with several local standard stars based on a calibration from the photometric night of oct . 21 with the nickel telescope . we present our low - resolution spectra of sn 2006jc in figure [ f : spec ] . in all spectra , we are able to identify intermediate - width , h@xmath7 , , , and emission lines . there are no photospheric p - cygni profiles typically found in early sn ib / c spectra @xcite ; the spectra seem to consist of two continuum components ( red and blue ) and intermediate - width ( @xmath8 ) emission lines . the redder he lines show a intermediate - width emission component ( fwhm @xmath9 ) similar to , but slightly wider than , that of sne iin @xcite . as seen in figure [ f : vel ] , the bluer he lines show narrow p - cygni profiles , with the absorption minimum blueshifted by roughly @xmath10 . the emission components of the lines with p - cygni profiles are narrower ( fwhm @xmath11 ) than the pure emission components , presumably because their blueshifted emission is self - absorbed . the only discernible characteristic between the groups of lines with the two different types of line profiles is wavelength ; in particular , the groups are not distinguished by singlet or triplet state . the line intensity ratios of the lines evolve with time , which may be an indication of changing densities or nlte effects . the spectra all show a relatively flat continuum redward of @xmath0 5500 , but a very steep blue continuum shortward of @xmath0 5500 , with a minimum @xmath12 color of @xmath13 mag . we have identified a few isolated lines in the spectra of sn 2006jc ( see figure [ f : spec ] ) , suggesting that there are many other lines , mostly at short wavelengths . it is likely that most of the apparent blue `` continuum '' consists of blended fe lines . @xcite suggested that emission lines create the excess blue continuum of sn 1988z . the blending of the fe lines makes disentangling the lines difficult , and detailed modeling is necessary to fully interpret our data . however , the amplitudes of the undulations in the blue continuum become larger with time , despite no significant change in the continuum shape . if the apparent continuum is the result of blended fe lines , the increase of these amplitudes follows the expected result of lines increasing relative to the continuum ( as seen with isolated lines ) . there are still many features that we are unable to identify , most notably that centered at 6357 which may be @xmath14 , with fwhm @xmath15 . this feature has some substructure and is wider than other emission lines , and thus is likely a blend of several lines . while most emission lines increase relative to the continuum with time , this feature decreases dramatically relative to the continuum and disappears by our nov . 23 spectrum , which is likely an excitation effect . since the lines increase relative to the continuum with time , it is most likely some other element . the features at 7881 ( somewhat blended with @xmath16 ) , 8215 and 9360 exhibit the same behavior and width , and may be of the same species . the slope of the blue continuum of sn 2006jc does not change much with time . the emission components of all he lines increase relative to the continuum with time , while some other species , including , decrease with time . the ir triplet also increases with time , consistent with other sne . the evolution of the blue continuum is apparent in figure [ f : lc ] . the @xmath17-band light curve declines more slowly than the other bands , which is unusual for a sn light curve . this supports the argument that the blue continuum is the result of a process not typically seen in sne . sn 2006jc has maximum absolute magnitudes of @xmath18 , @xmath19 , @xmath18 , and @xmath20 for , respectively , making it as luminous as sn 1994i @xcite . assuming that nearly all of its luminosity comes from @xmath21ni to @xmath21co to @xmath21fe decay , the fact that sne 1994i and 2006jc have similar luminosities suggests they created similar @xmath21ni masses , @xmath0 0.07 @xmath22 @xcite . the early - time decline rates of sne 2006jc , 2002ao , and 1999cq are much faster than those of normal sne iin ( see figure [ f : lc ] for a comparison to sn iin 2002bu ) . the early light curves of sne iin are powered mainly by circumstellar interaction and their slow declines at early times indicate a massive csm . despite showing obvious features of circumstellar interaction in their spectra , sne 2006jc , 2002ao , and 1999cq all decline very quickly , suggesting an extremely dense , but low - mass and clumpy csm . the late - time decline rates of these objects are also very fast , indicating either a small @xmath21ni mass or a small ejecta mass . the strength of the intermediate - width emission lines in sn 2006jc relative to other sne ib is indicative of a dense csm rich in he . similar to the h lines seen in sn iin spectra , the he lines of sn 2006jc are the result of the sn ejecta interacting with a dense csm . the lines seen only in emission have fwhm @xmath23 20003000 km s@xmath4 , and the absorption minimum of the p - cygni profiles is @xmath0 1000 km s@xmath4 , velocities typical of wr winds @xcite . the identification of h@xmath7 suggests that the progenitor recently ejected material containing at least some h into its surroundings , perhaps meaning that the progenitor recently evolved from a lbv to a wne wr star @xcite . additionally , the smaller width of the h@xmath7 line is consistent with a different ejection mechanism from that of the he . lbvs typically have slower expansion speeds than wr stars , indicating that the h might have been ejected during the lbv stage , while the he was ejected during the wr stage . @xcite did not see any h@xmath7 in the spectrum of sn 1999cq , but the low signal - to - noise ratio of the spectrum makes the measurement ambiguous . the spectrum of sn 2002ao does show relatively strong h@xmath7 ( see figure [ f : spec ] ) , suggesting that its csm contains more h than that of sn 2006jc . as noted by @xcite , there was a bright ( @xmath24 mag ) outburst coincident with the position of sn 2006jc in oct . 2004 , which was initially thought to be an lbv outburst similar to sn 1961v @xcite . unless it was a mere coincidence , it seems that the progenitor of sn 2006jc suffered an event analogous to the non - terminal eruptions of lbvs shortly before final core collapse , ejecting a he - rich shell with which the sn ejecta are now interacting . the short time between these events may have far - reaching implications for the late evolution of massive stars , beyond what we can discuss here ; however , it is worth noting that @xcite determined that sn iin 1994w had an outburst @xmath0 1.5 years prior to explosion . equally surprising is that this type of event may have happened in a star inferred to be a wn star ; while such outbursts are known to occur in lbvs that still have their h envelopes , no such variability has been documented in wn stars . it may hint that the progenitor of sn 2006jc had just recently transitioned from the lbv to the wn stage , supporting the notion that lbv mass loss facilitates the onset of the wr phase . this would seem consistent with the presence of h in its csm . the slow @xmath17-band decline relative to the @xmath25 band , the steep blue continuum , and the p - cygni profiles for all suggest that the emission at wavelengths shorter than @xmath0 5500 arises from a mechanism that is different from that which produces the red continuum . we identified some isolated lines , suggesting that many other lines are present in the spectrum , but are blended and appear as a pseudo - continuum because of their large line widths . since we do not see other high - excitation lines , this emission might be provided by fluorescence and not collisional excitation . @xcite noted that sn 1999cq had a bluer continuum than normal sne ic . they also noted that sn 1999cq had @xmath26 mag , and argued that @xmath27 mag . similarly , sn 2002ao may be extinguished . correcting for reasonably small reddening , the continua of both sne 1999cq and 2002ao have similar colors to that of sn 2006jc , indicating that these events may also have a strong blue continuum . the main difference between the spectra is the strong h&k absorption in sn 1999cq . the double - peaked nature of some lines suggest a complex , asymmetric csm . unlike sne iin , the light curve of sn 2006jc declined very quickly , suggesting that there is little mass in the csm . however , the strong emission lines indicate a very dense csm . these facts are compatible if the csm is asymmetric or clumpy . also , the p - cygni profiles seen only for the blue lines also indicate asymmetry , because the blue continuum is absorbed by the he gas , whereas the red continuum ( presumably from the inner sn debris ) is not . a simple way to explain these observables is supernova ejecta interacting with a bipolar csm , which causes the double - peaked nature of the he lines . the he - rich csm , which was likely the result of the recent ( 2004 ) outburst , may be asymmetric because of either rapid rotation in the progenitor or strong interaction with a binary companion . some of the data presented herein were obtained at the w. m. keck observatory , which was made possible by the generous financial support of the w. m. keck foundation . we thank the keck and lick observatory staffs for their assistance . this research was supported by nsf grant ast-0607485 and the tabasgo foundation . kait was made possible by generous donations from sun microsystems , inc . , the hewlett - packard company , autoscope corporation , lick observatory , the nsf , the university of california , and the sylvia & jim katzman foundation .

we present optical photometry and spectra of the peculiar type ib supernova ( sn ) 2006jc . strong and relatively narrow emission lines indicate the progenitor star exploded inside a dense circumstellar medium ( csm ) rich in he . an exceptionally blue apparent continuum persists from our first spectrum obtained 15 days after discovery through our last spectrum @xmath0 1 month later . based on the presence of isolated emission lines , we interpret the blue `` continuum '' as blended , perhaps fluorescent , fe emission . one or two of the reddest line profiles in our spectra are double peaked , suggesting that the csm has an aspherical geometry . the lines that are superposed on the blue continuum show p - cygni profiles , while the redder lines do not , implying that the blue continuum also originates from an asymmetric mass distribution . the he - rich csm , aspherical geometry , and line velocities indicate that the progenitor star was a wne wolf - rayet ( wr ) star . a recent ( 2 years before the sn ) , coincident , luminous outburst similar to those seen in luminous blue variables ( lbvs ) is the leading candidate for the dense csm . such an eruption associated with a wr star has not been seen before , indicating that the progenitor star may have recently transitioned from the lbv phase . we also present unpublished spectral and photometric data on sn 2002ao which , along with sn 1999cq , is very similar to sn 2006jc . we propose that these three objects may represent a new and distinct class of sne arising from wr progenitors surrounded by a dense csm .

the stimulated raman transition is an extremely powerful tool for laser manipulation of cold atoms and ions . by coupling long - lived states via , but never populating , radiative states , experimenters can emulate near - ideal two - level quantum systems with no significant decay @xcite . this technique has been used to measure sub - linewidth features @xcite and to construct atomic interferometers which , by exploiting photon recoil , create spatially separated atomic wave packets which are sensitive to gravity @xcite or fundamental constants @xcite . the effective two - level system , which emerges from the raman problem , can exhibit behavior such as rabi flopping @xcite , can be used for experiments such as ramsey interferometry @xcite , and can provide the qubits for quantum information processing @xcite . sequences of raman pulses can be used to craft arbitrary superpositions in systems with numerous metastable states @xcite and to prepare such systems in particular states prior to coherent manipulation @xcite . raman processes have also been used to cool atomic samples to far below the photon recoil limit @xcite . throughout the literature , when the raman transition is discussed , the level structure of the atom is often approximated to three levels two metastable states and one intermediate ( radiative ) state . the raman problem is solved for this prototypical case and then extended , without proof , to include the multilevel structure of the atom by summing over the various possible routes ( see , e.g. , ref . @xcite ) . here , by including multiple routes from the outset , we confirm that this simple approach is correct , show how an expression for the lightshift emerges naturally from this treatment , and show that the system behaves as a two - level system with an effective coupling strength and an effective detuning . there is much existing work related to this problem . the three - level ( single intermediate state ) off - resonant case has been treated @xcite , there have been extensions to four levels @xcite , and the general multilevel problem has been recast into `` serial '' and `` parallel '' cases @xcite . it has never been shown rigorously , however , that the three - level case can be extended in the way so often assumed . in the following , we use the semiclassical approach but alternatively one might consider the jaynes cummings model @xcite . this article is structured as follows . we first describe , in , the raman transition in a three - level system and we show how this can be generalized to include multiple intermediate states ; details of the lengthly calculation are confined to the appendix . we then derive , in , expressions for the behavior of the quantum - mechanical amplitudes in the general , off - resonance case . in we show how these results can be applied to alkali - metal atoms , and we conclude in . the simplest system in which a raman transition may be driven is the three - level `` @xmath0 '' system , illustrated in , in which two long - lived ground states are coupled via a radiative upper state which , because the single - photon detuning is sufficiently large , is never significantly populated . a simple three - level `` @xmath0 '' system , in which a raman transition between states @xmath1 and @xmath2 is driven by `` pump '' and `` stokes '' fields via intermediate state @xmath3 . for each field , the coupling strength and frequency are shown in parentheses , and frequencies are chosen to be near two - photon resonance : @xmath4 ; @xmath5 . the single - photon detuning @xmath6 is large compared with the couplings , @xmath7 , and , in this illustration , is negative : @xmath8 . the two - photon detuning @xmath9 is small compared with the separation between the ground states and , in this illustration , is also negative.,width=302 ] we label the states of the system by @xmath10 , with states @xmath1 and @xmath3 coupled by the `` pump '' field of strength @xmath11 and frequency @xmath12 , and states @xmath2 and @xmath3 coupled by the `` stokes '' field of strength @xmath13 and frequency @xmath14 . using the usual correspondence between bra ket and column vector notation ( see , e.g. , ref . * ii - c ) ) , the hamiltonian for this system may be represented by the following matrix ( see , e.g. , ref . @xcite ) : @xmath15here , as in standard treatments , we assume there is no coupling between states @xmath1 and @xmath3 by the stokes field , or between states @xmath2 and @xmath3 by the pump field . this hamiltonian can be simplified by making the rotating wave approximation and transforming to the interaction picture . this yields a slowly varying hamiltonian , @xmath16 where the pump frequency @xmath17 is detuned from single - photon resonance by @xmath6 , and the difference between the pump frequency and the stokes frequency @xmath18 is offset by @xmath9 from the two - photon resonance @xmath19 . we now extend this interaction picture to include multiple levels and define the hamiltonian @xmath20 to be @xmath21 the second part of the subscript , @xmath22 , denotes the level to which the pump or stokes field couples . note that the oscillation frequency is the same for each stokes term because this depends on the difference in the frequency of the fields and not on the bohr energy of the intermediate level . however , the single - photon detunings do depend on the intermediate level bohr frequencies , but we now make the approximation that the detuning is large compared with the separation of these intermediate levels , and hence @xmath23 . in this limit , the above hamiltonian describes a raman system and we expect to see oscillations of population between the ground states . we solve the schrdinger equation with this hamiltonian by using unitary transformations to find a basis where the time evolution of the states is simple and the transformed hamiltonian is diagonal . when one makes such a conversion between bases , is it possible to find an equivalent schrdinger equation with a transformed hamiltonian @xcite : if @xmath24 , then @xmath25 where @xmath26 for the multistate hamiltonian @xmath20 in , we choose the operator @xmath27 to be the matrix of eigenvectors ; the first term @xmath28 is thus the diagonal matrix of the eigenvalues . in the appendix , we detail a procedure to find the eigensystem of this hamiltonian ; we find two eigenvectors which are superpositions of these ground states , and @xmath29 more which are superpositions of the remaining intermediate levels . these @xmath29 upper states are decoupled from the ground states and so we ignore them in the following treatment . the difference between the eigenvalues for the two ground state eigenvectors is @xmath30 where , for conciseness , we have represented the couplings as vectors @xmath31 and @xmath32 with components @xmath33 and @xmath34 respectively , and have used vector notation for dot products and norms . this oscillation frequency is composed of a coupling strength @xmath35 and a detuning @xmath36 , via @xmath37 , analogously to a two - level system : @xmath38 the detuning @xmath36 is readily identified as the lightshift and we justify this in the next section . the operator which describes the transformation from the bare ground states to these dressed ground states can , because of normalization , be written as a rotation : @xmath39 and the angle @xmath40 is defined by @xmath41 . this treatment is sufficient for the on two - photon resonance case , where @xmath42 , but in general the effective hamiltonian @xmath43 also contains a time - derivative second term , originating from the time - dependence of the operator @xmath27 . away from the two - photon resonance , where @xmath44 , we find the following slowly varying , but nevertheless time - dependent , effective hamiltonian : @xmath45 the hamiltonian @xmath43 in has the same form as that for the simple two - level problem in the interaction picture and with the rotating - wave approximation . we can , therefore , use familiar tools to solve this problem . first , we transform to find a time - independent hamiltonian using the operator @xmath46 in : @xmath47 next , analogously to the dressed - states approach , we rotate by an angle @xmath48 ( thus defining @xmath49 ) where @xmath50 to find a diagonal hamiltonian @xmath51 . the difference between the diagonal elements of @xmath51 corresponds to the phase evolution frequency @xmath52 of the states in this basis . as in the previous section , we see that this oscillation frequency is composed of a coupling strength @xmath35 and a modified effective detuning @xmath53 : @xmath54 , where @xmath55 is the detuning relative to @xmath36 , which was previously identified as the lightshift . we now relate the pure phase evolution in this doubly dressed basis to the evolution of the bare states by concatenating the transformations that led us to this final hamiltonian : @xmath56 and @xmath57 , or @xmath58 where @xmath59 and @xmath60 are the ground- and excited - state components of the doubly dressed wave function @xmath61 and the bare ( interaction picture ) wave function @xmath62 , respectively . finally , the time evolution of the doubly dressed states is simply @xmath63 using we can find the dressed state initial conditions @xmath64 in terms of the bare state initial conditions @xmath65 . using these values , we can then use to find the dressed state coefficients at some later time . finally , we can invert the transformation in to find the time evolution of the bare state amplitudes . the time dependence of the bare state coefficients @xmath66 is readily calculable from the procedure described above and is stated here for completeness : [ eqn : explicita01 t ] @xmath67\nonumber\\ & \phantom{=}\quad~~+a_1(0)\ , i{\frac{\omega_\text{d}}{\widetilde{\omega}_\text{d}}}\sin{\left(\tfrac{1}{2}\widetilde{\omega}_\text{d } t\right)}\biggr)\,;\phantom{e^{-i\delta t}}\\ a_1(t)&=\biggl(a_1(0)\left[\cos{\left(\tfrac{1}{2}\widetilde{\omega}_\text{d } t\right)}+i{\frac{\delta_\text{d}}{\widetilde{\omega}_\text{d}}}\sin{\left(\tfrac{1}{2}\widetilde{\omega}_\text{d } t\right)}\right]\nonumber\\ & \phantom{=}\quad~~+a_0(0)\ , i{\frac{\omega_\text{d}}{\widetilde{\omega}_\text{d}}}\sin{\left(\tfrac{1}{2}\widetilde{\omega}_\text{d } t\right)}\biggr)e^{-i\delta t}\,.\end{aligned}\ ] ] hence the system behaves as a two - level system with the coupling strength @xmath68 and detuning @xmath55 , relative to the effective detuning @xmath36 . this justifies our previous identification of @xmath36 with the lightshift . the complete , but cumbersome , formulas in describe the behavior of the bare - state amplitudes in terms of the bare state initial conditions . if , instead , we express this evolution in terms of the initial values in the doubly dressed basis , we see clearly that the evolution is composed of a time - independent offset and an oscillation : @xmath69 the population @xmath70 in state @xmath1 therefore oscillates with peak - to - peak amplitude no greater than @xmath71 $ ] which , expressed in terms of the effective coupling strength and detuning , is @xmath72 this envelope function describes a power - broadened lorentzian , centered on the light shifted frequency difference between the ground states . this expression represents the maximum possible population transfer , and any oscillation will be contained within this envelope . a few specific cases are provided here for illustration . first , for @xmath73 and @xmath74 we recover the well - known results for the three level problem . next we note two interesting cases : for @xmath75 , the lightshift @xmath36 is zero . on the other hand , for @xmath76 , the rabi frequency @xmath35 is zero . if , as in this last case , the coupling vectors @xmath31 and @xmath77 are orthogonal , then the transition is not driven . examples include the trivial case where there is no intermediate state to which both ground states are coupled and the case where there are states to which both are coupled , but where these individual coupling strengths sum to zero . however , unless the vectors are orthogonal , it is possible to adjust the pump and stokes field strengths to ensure the norms of the vectors are equal , and hence that the lightshift is zero . our approach relies on the slow time - dependence of the interaction - picture hamiltonian : we require that the system is near two - photon resonance , as previously stated , and that there is no coupling of state @xmath1 by the stokes field or of state @xmath2 by the pump field . if present , these cross - coupling terms would cause the off - diagonal terms in @xmath20 to oscillate in amplitude as well as phase , and the treatment in the appendix would no longer be valid . for our treatment to be valid , it must therefore be possible to identify clearly which field is resonant with which transition ( see refs @xcite and @xcite for further discussion ) . first , the detuning must be such that no field is close to a single - photon resonance : @xmath78 and @xmath79 . additionally , the coupling strength must be sufficiently small that each ground state can be resolved : @xmath80 . if this last condition is violated , the system may still appear ramanlike and exhibit coherent behavior , but it is not described adequately by the treatment in this article . we are able to calculate the rabi frequency and lightshift for two states @xmath1 and @xmath2 coupled via a number of upper states . a common embodiment of this situation is the coupling of two ground hyperfine states via a manifold of upper hyperfine states in an alkali - metal atom . indeed , there typically exists many such pairs of states , but , as ensured by conservation of angular momentum , one state is raman coupled to at most one other state ; hence , the total system may be treated as a collection of independent pairwise couplings . as a typical example , consider the raman transition between the ground hyperfine states , via the radiative upper states , in atomic cesium . the pump and stokes fields , both tuned near to the @xmath81 transition at @xmath82 , couple states @xmath83 and @xmath84 , respectively , to the @xmath85 manifold , and have a frequency difference near to the hyperfine splitting of @xmath86 @xcite . the two ground states and the intermediate states are @xmath87 where @xmath88 are the polarizations of the co - propagating pump and stokes fields and correspond to linear and left and right circular polarizations , respectively ; @xmath89 labels the zeeman sublevel , corresponding to the projection of the total angular momentum @xmath90 along the quantization axis provided by an external magnetic field @xcite . linear polarization , in this context , refers specifically to the case of the light electric field parallel to the quantization axis ; if these axes are orthogonal , the light field interacts with the atom as though it were a superposition of left and right circular polarizations . the coupling strengths @xmath91 depend not only on the light intensities but also on the dipole matrix element for the transition . using the wigner - eckart theorem @xcite we can split the overlap integral needed to find this dipole matrix element and extract from it a purely geometrical term @xmath92 , leaving a term which embodies the other physical details of the transition : @xmath93 the `` reduced '' matrix element , denoted by double bars @xmath94 , depends on many details of the atom , including nuclear mass , and is not easily calculated ; it can , however , be found experimentally from measurements of the upper - state lifetime , as described by loudon ( * ? ? ? * eq . ( 2.57 ) ) and demtrder @xcite : @xmath95 the second part is the product of geometrical terms : @xmath96 where the arraylike symbol in parentheses @xmath97 is the wigner 3-@xmath98 symbol and the similar term in braces @xmath99 is the wigner 6-@xmath98 symbol @xcite ; both are closely related to the clebsch - gordan coefficients . this relation is described in detail by edmonds @xcite . the two states coupled by the raman interaction are both in @xmath100 so , in the calculations that follow , it is only this geometrical term which is relevant . we imagine the atom in a region of uniform magnetic field and consider an experiment where it is possible to adjust the frequency difference in order to sweep across transitions between various zeeman sublevels . the properties which affect the dipole moment are the various quantum numbers : the nuclear spin @xmath101 ; the electron angular momentum @xmath102 or @xmath103 ; the total angular momentum @xmath104 to @xmath105 via @xmath106 ; and the aforementioned projection @xmath89 of @xmath90 along the quantization axis . the coupling strength for a dipole transition between states @xmath10 and @xmath107 is proportional to the electric field @xcite : @xmath108 where @xmath109 is the dipole operator . as above , we can extract a geometrical term and , using the vector notation , @xmath110 hence , the relative properties of each of the zeeman sublevels are determined by the geometrical terms @xmath111 . if we revisit the equations for the coupling strength ( ) and the lightshift ( ) , we see that these terms appear as @xmath112 and @xmath113 , respectively . thus : @xmath114 it is simple to calculate these factors for a given initial state and pair of polarizations to examine how the lightshift and the coupling strength depend on the strength of the applied fields . we find that @xmath115 , for transitions from @xmath116 to @xmath117 , driven by light with polarization @xmath118 , are , for an alkali - metal atom with nuclear spin @xmath119 , given by @xmath120 where @xmath121 for the @xmath122 transition and @xmath123 for the @xmath81 transition . the pump vector @xmath124 couples from the lower hyperfine state and corresponds to the negative branch ( @xmath125 ) ; the stokes vector @xmath126 couples from the upper hyperfine state and corresponds to the positive branch ( @xmath127 ) . using these expressions and , one may easily calculate the lightshift for a given raman transition in any alkali - metal atom . we now turn to the coupling strength @xmath35 which , for equal polarizations @xmath128 , is symmetrical about @xmath129 . we find @xmath130 for the @xmath122 and @xmath81 transitions , where @xmath131 is given above . if we now break this symmetry by choosing , for example , @xmath132 , we find @xmath133 where @xmath134 is the @xmath135 triangular number ( @xmath136 ) . the dependence of coupling strength on @xmath137 level is illustrated in . .scaling of the coupling strength and the lightshift for the transition @xmath138 to @xmath139 for @xmath128 , in cesium , in terms of the geometrical parts of the dipole matrix elements , as described by eqs ( [ eqn : niceg - squared ] ) and ( [ eqn : nicegpgs11 ] ) . [ cols="^,>,>,>",options="header " , ] the values for a common arrangement are shown in . while of course linear in any overall scaling of the intensity , the lightshift has a different dependence on the individual field strengths for each of the zeeman sublevels . it is offset from zero ( for unequal intensities ) and is linear in @xmath89 ; the lightshift between the hyperfine ground states hence has the same dependence on @xmath89 as the zeeman shift , and , in , e.g. , ref . @xcite , is sufficient to account for the majority of the observed spacing between the spectral peaks . the coupling strengths are not necessarily symmetrical about @xmath129 and , for an experiment in which damping is important , this may be manifest as a change in the amplitude of the peaks . however , the common arrangement of equal polarizations does not show this asymmetry , and , as noted in ref . @xcite , asymmetry in the preparation of the initial states is also important . we have described the stimulated raman transition and , by including multiple intermediate states from the outset , have obtained results which give the coupling strength for the multistate system and from which the lightshift naturally emerges . we have applied this method to the cesium atom , given more general expressions for the alkali - metal atoms , and noted how the linear dependence of the lightshift on the @xmath89 level mimics a zeeman shift . we comment on the possibility of the dependence of the coupling strength on @xmath89 level manifesting as a variation in peak height in experimental spectra . our results were derived in the limit of far detuning and our calculations were simplified greatly by this enforced absence of decoherence . however , for specific coupling strengths and detunings , the problem of finding eigenvalues and vectors can be treated numerically , and many efficient algorithms exist for this task . hence , a similar approach might be used for situations including coherent population trapping @xcite and electromagnetically induced transparency @xcite . we thank steven barnett and dieter meschede and his group for helpful discussions . this work was supported by the uk engineering and physical sciences research council ( epsrc ) grants ep / e058949/1 and ep / e039839/1 , and the european science foundation s euroquam project _ cavity - mediated molecular cooling_. in this appendix , we find the eigenvalues and eigenvectors for the matrix @xmath140 representing the hamiltonian @xmath20 of our multilevel system , as described in sec . [ sec : three_and_multi_level_systems ] . for brevity in the derivation , we make the replacements @xmath141 , @xmath142 , and @xmath143 . the frequency @xmath144 is , conceptually , the natural frequency scale for the problem . we calculate the determinant of the @xmath145 matrix @xmath146 , @xmath147 to find the characteristic equation and hence the eigenvalues @xmath148 of @xmath140 . if we define @xmath149 to be the matrix @xmath150 with row @xmath151 and column @xmath98 removed , and @xmath152 to be the element @xmath153 of matrix @xmath150 , then using expansion by minors , @xmath154 we first evaluate the term @xmath155 : @xmath156 as before , we decompose the determinant in terms of the elements in the first column . the first minor matrix is diagonal @xmath157 , and hence the first term is @xmath158 . for subsequent terms @xmath159 ; @xmath160 , we find @xmath161 , where the problem of calculating the determinant of each minor matrix is greatly simplified by swapping columns to ensure each is upper diagonal with diagonal elements @xmath162 . overall , these terms sum to @xmath163 . hence , @xmath164 we now consider the terms @xmath165 for @xmath160 : @xmath166 before deconstructing @xmath167 , we note that the next term , @xmath168 , is of the same form if we interchange the columns headed @xmath169 and @xmath170 : @xmath171with a sign change . the first term of this determinant has prefactor zero and so can be ignored . for the second term , the minor matrix @xmath172 is upper diagonal and hence has determinant @xmath173 and , similarly , @xmath174 for @xmath160 . we now investigate the determinants @xmath175 for @xmath176 . in the expansion of @xmath177 , the only terms which have nonzero coefficient are @xmath178 and @xmath179 . we can extend this treatment for @xmath160 and @xmath176 . finally , we find @xmath180 the equation @xmath181 clearly has solution @xmath182 with multiplicity @xmath183 . with this factor removed , and using @xmath184 to phrase this equation in terms of vectors , we obtain @xmath185 which is a fourth - order polynomial in @xmath148 ( with leading coefficient @xmath186 ) and hence the product of the remaining roots @xmath187 equals the constant term . this term is finite , and so _ at least _ one @xmath187 becomes negligible as @xmath188 , and hence we can make the approximation @xmath189 in this limit . the resulting equation is a quadratic in @xmath148 with solutions @xmath190 the eigenvalue equation @xmath191 yields the following : @xmath192 if we multiply by @xmath193 and sum over @xmath194 , and then enforce the normalization condition @xmath195 , we arrive at @xmath196=\lambda\left(|a_0|^2 + |a_1|^2\right)\,\text{,}\ ] ] and hence @xmath197 which , in the limit of large @xmath9 , tends to unity . by way of confirmation , we see from that @xmath198 which tend to zero in this limit . the two eigenstates associated with the two finite eigenvalues are hence orthogonal superpositions of the two ground eigenstates ; we represent this transformation as a rotation and proceed to find its angle . using eqs ( [ eqn : eigenvector1 ] ) and ( [ eqn : eigenvector3 ] ) , we have @xmath199\nonumber\\ & = & \frac{1}{\lambda(\lambda-\delta)}\left[a_0{\boldsymbol{x}}\cdot{\boldsymbol{x}}^ * + a_1 { \boldsymbol{x}}\cdot{\boldsymbol{y}}^*\right]\,\text{,}\end{aligned}\ ] ] and similarly for @xmath200 . hence we obtain @xmath201 which , because @xmath202 , we can express as the tangent of an angle : @xmath203 where @xmath204 and @xmath205 . the transformation from the bare - state basis to this dressed - state basis can hence be described by the rotation @xmath206 and oscillations are thus driven with amplitude @xmath207 at the rate @xmath208

we consider the stimulated raman transition between two long - lived states via multiple intermediate states , such as between hyperfine ground states in the alkali - metal atoms . we present a concise treatment of the general , multilevel , off - resonant case , and we show how the lightshift emerges naturally in this approach . we illustrate our results by application to alkali - metal atoms and we make specific reference to cesium . we comment on some artifacts , due solely to the geometrical overlap of states , which are relevant to existing experiments .

in this work we study rotating waves in rings of neurons described by the theta model . the theta model @xcite , which is derived as a canonical model for neurons near a ` saddle - node on a limit cycle ' bifurcation , assumes the state of the neuron is given by an angle @xmath0 , with @xmath1 , @xmath2 corresponding to the ` firing ' state , and the dynamics described by @xmath3 where @xmath4 represents the inputs to the neuron . when @xmath5 this represents an ` excitable ' neuron , which in the absence of external input ( @xmath6 ) approaches a rest state , while if @xmath7 this represents an ` oscillatory ' neuron which performs spontaneous oscillations in the absence of external input . a model of synaptically connected neurons on a continuous spacial domain @xmath8 takes the form : @xmath9,\ ] ] @xmath10 where @xmath11 is a positive function and @xmath12 is defined by @xmath13 here @xmath14 ( @xmath15 ) measures the synaptic transmission from the neuron located at @xmath16 , and according to ( [ syn]),([defp ] ) it decays exponentially , except when the neuron fires ( _ i.e. _ when @xmath17 , @xmath2 ) , when it experiences a jump . ( [ de ] ) says that the neurons are modelled as theta - neurons , where the input @xmath18 to the neuron at @xmath16 , as in ( [ thn ] ) , is given by @xmath19 @xmath20 ( here assumed to be positive ) describes the relative strength of the synaptic coupling from the neuron at @xmath16 to the neuron at @xmath21 , while @xmath22 is a parameter measuring the overall coupling strength . the above model , in the case @xmath23 , is the one presented in @xcite . in the case @xmath24 this model is the one presented in @xcite ( remark 2 ) and @xcite . we always assume @xmath25 . when the geometry is linear , @xmath26 , it is natural to seek travelling waves of activity along the line in which each neuron makes one or more oscillations and then approaches rest . in @xcite it was proven that for sufficiently strong synaptic coupling @xmath27 , at least two such waves , a slow and a fast one , exist , and also that they always involve each neuron firing more than one time before it approaches rest , while for sufficiently small @xmath27 such waves do not exist . it was not determined how many times each neuron fires before coming to rest , and it may even be that each neuron fires infinitely many times . some numerical results in the case of a one and a two - dimensional geometry were obtained in @xcite . in this work we consider a different possibility for the spacial geometry : @xmath28 , so the neurons are placed on a ring and our equations are ( [ syn ] ) and @xmath29 with @xmath30 where @xmath11 is continuous , positive and periodic @xmath31 @xmath32 and the solutions satisfy the periodicity conditions @xmath33 @xmath34 the integer @xmath35 ( the ` winding number ' ) is determined by the initial condition @xmath36 , and will be preserved as long as the solution remains continuous . in this geometry , a different kind of wave of activity is possible : a wave that rotates around the ring repeatedly . such waves , that is solutions of the form : @xmath37 @xmath38 where @xmath39 is the wave velocity , are the focus of our investigation . in section [ pre ] we show that in the case that the winding number @xmath40 , there can exist only trivial rotating waves . thus the interesting cases are when @xmath41 . here we study the case @xmath42 , the case @xmath43 being beyond our reach . thus , this work concentrates on the first non - trivial case . our central results about existence , nonexistence and multiplicity of rotating waves can be summarized as follows ( see figures [ syn1],[syn2 ] for the _ simplest _ diagrams consistent with these results ) : [ sum ] consider the equations ( [ ge]),([syn ] ) with conditions ( [ bnt ] ) , ( [ bns ] ) , and @xmath42 . \(i ) in the oscillatory case @xmath7 : for all @xmath22 there exists a rotating wave , with velocity going to @xmath44 as @xmath45 . \(ii ) in the excitable case @xmath5 : \(i ) for @xmath22 sufficiently small there exist no rotating waves . \(ii ) for @xmath27 sufficiently large there exist at least two rotating waves , a ` fast ' and a ` slow ' one , in the sense that their velocities approach @xmath44 and @xmath46 , respectively , as @xmath47 . therefore our results bear resemblance to those obtained in @xcite for the case of a linear geometry . we note that although for the rotating waves found here each neuron fires infinitely many times , the reason for this is that it is re - excited each time , because of the periodic geometry . during each revolution of the rotating wave , each neuron fires once , so naively one could think that the analogous phenomenon in a linear geometry would be a travelling wave with each neuron firing once - but this was shown to be impossible in @xcite . it is interesting to note that while in @xcite some restrictions were made on the coupling function @xmath11 , like being decreasing with distance , here no such restrictions are imposed beyond ( [ jpos ] ) , ( [ jper ] ) . we would expect however that some restriction would need to be imposed on @xmath11 in order to obtain stability of travelling waves . the whole issue of stability remains quite open and awaits future investigation . in the case @xmath26 , both numerical evidence in @xcite and results obtained in other models @xcite indicate that the fast wave is stable while the slow wave is unstable , so we might conjecture that this is true for the case investigated here as well - at least under some natural assumptions on @xmath11 . some analytical progress on the stability question in the case @xmath26 has recently been achieved in @xcite . let us note that the model considered here , in the case @xmath5 , describes waves in an excitable medium , about which an extensive literature exists ( see @xcite and references therein ) . however , most models consider diffusive rather than synaptic coupling . in the case of the theta model on a ring , with _ diffusive _ coupling , and @xmath42 , it is proven in @xcite that a rotating wave exists regardless of the strength of coupling ( _ i.e. _ the diffusion coefficient ) , so that our results highlight the difference between diffusive and synaptic coupling . in section [ reduction ] we reduce the study of rotating waves to the investigation of the zeroes of a function of one variable . in section [ constant ] we investigate the special case in which the coupling is uniform ( @xmath48 is a constant function ) , which , although artificial from a biological point of view , allows us to obtain closed analytic expressions for the wave - velocity vs. coupling - strength curves in an elementary fashion . we can thus gain some intuition for the general case , and obtain information which is unavailable in the case of general @xmath11 , like precise multiplicity results . it is interesting to investigate to what extent the more precise results obtained in the uniform - coupling case extend to the general case , and we shall indicate several questions , which remain open , in this direction . in section [ general ] we turn to the case of general coupling functions @xmath11 , and prove the results of theorem [ sum ] above , obtaining also some quantitative estimates : lower and upper bounds for the critical values of synaptic coupling coupling strength @xmath27 , as well as for the wave velocities . we begin with an elementary calculus lemma which is useful in several of our arguments below . [ cut ] let @xmath49 be a differentiable function , and let @xmath50 , @xmath51 , be constants such that we have the following property : @xmath52 then the equation @xmath53 has at most one solution . assume by way of contradiction that the equation @xmath53 has at least two solutions @xmath54 . define @xmath55 by @xmath56 @xmath57 is nonempty because @xmath58 . let @xmath59 . by continuity of @xmath60 we have @xmath61 . we have either @xmath62 or @xmath63 , and we shall show that both of these possibilities lead to contradictions . if @xmath62 , then by ( [ kp ] ) we have @xmath64 @xmath65 so we conclude that there exists @xmath66 with @xmath67 , contradicting the definition of @xmath68 . if @xmath63 then @xmath69 is a limit - point of @xmath57 , which implies that @xmath70 , contradicting ( [ kp ] ) . these contradictions conclude our proof . turning now to our investigation , we note a few properties of the functions @xmath71 and @xmath72 defined by ( [ spec ] ) which will be used often in our arguments : @xmath73 @xmath74 @xmath75 @xmath76 plugging ( [ tw1]),([tw2 ] ) into ( [ ge ] ) , ( [ syn ] ) , and setting @xmath77 we obtain the following equations for @xmath78 , @xmath79 : @xmath80 @xmath81 in order to satisfy the boundary conditions ( [ bnt]),([bns ] ) , @xmath82 and @xmath83 have to satisfy @xmath84 let us first dispose of the case of zero - velocity waves , @xmath86 . we get the equations @xmath87 @xmath88 if there exists some @xmath89 with @xmath90 , @xmath2 , then , substituting @xmath91 into ( [ teq00 ] ) and using ( [ hpi]),([wz ] ) , we obtain @xmath92 , a contradiction . hence we must have @xmath93 which implies that @xmath94 , so that ( [ req00 ] ) gives @xmath95 , and ( [ teq00 ] ) reduces to @xmath96 , and thus @xmath78 is a constant function , the constant being a root of @xmath71 . this implies , first of all , that the winding number @xmath35 is @xmath46 , since a constant @xmath78 can not satisfy ( [ bnp0 ] ) otherwise . in addition the function @xmath71 must vanish somewhere , which is equivalent to the condition @xmath97 . we have thus proven [ zvel ] zero - velocity waves exist if and only if @xmath40 and @xmath97 , and in this case they are just the stationary solutions @xmath98 @xmath99 we will now show that the trivial waves " of lemma [ zvel ] are the only ones that occur for @xmath40 . [ mz ] assume @xmath40 . \(i ) if @xmath7 there are no rotating waves . \(ii ) if @xmath97 the only rotating waves are those given by lemma [ zvel ] . assume @xmath100 is a solution of ( [ teq0]),([req0 ] ) satisfying ( [ bnp0 ] ) with @xmath40 , _ i.e. _ @xmath101 and ( [ bnr0 ] ) . we also assume @xmath102 , otherwise we are back to lemma [ zvel ] . we shall prove below that @xmath78 must satisfy ( [ npi ] ) , and hence that @xmath103 , so that by ( [ req0]),([bnr0 ] ) we have @xmath95 , so that ( [ teq0 ] ) reduces to @xmath104 since @xmath102 , if @xmath71 has no roots ( @xmath7 ) , ( [ red ] ) has no solutions satisfying ( [ bnp00 ] ) . if @xmath71 does have roots ( @xmath97 ) then the only solutions of ( [ red ] ) satisfying ( [ bnp00 ] ) are constant functions , the constant being a root of @xmath71 , and we are back to the same solutions given in lemma [ zvel ] , which indeed can be considered as rotating waves with arbitrary velocity . it remains then to prove that ( [ npi ] ) must hold . assume by way of contradiction that @xmath90 for some integer @xmath105 . by ( [ hpi]),([wz]),([teq0 ] ) , and the assumption @xmath106 , we have @xmath107 thus the assumptions of lemma [ cut ] , with @xmath108 , @xmath109 , @xmath110 , are satisfied , and we conclude that the equation @xmath111 has at most one solution , contradicting the fact that , by ( [ bnp00 ] ) , we have @xmath112 . having found all possible rotating waves in the case @xmath40 , we can now turn to the case @xmath41 . in fact , as was mentioned in the introduction , we shall treat the case @xmath42 , the cases @xmath43 being harder . by lemma [ zvel ] we know that there are no zero - velocity waves , so we can assume @xmath102 and define @xmath113 so that our equations for the rotating waves can be rewritten @xmath114 @xmath115 with periodic conditions @xmath116 @xmath117 we study the equations ( [ teq]),([req ] ) for @xmath118 with periodic conditions ( [ bnp]),([bnr ] ) . we will derive a scalar equation ( see ( [ tt ] ) below ) so that rotating waves are in one - to - one correspondence with solutions of that equation . we note first that , since by ( [ bnr ] ) we have @xmath119 , and since any rotating wave generates a family of other rotating waves by translations , we may , without loss of generality , fix @xmath120 the following lemma shows that for a rotating wave ( in the case @xmath42 ) there is at any specific time a unique neuron on the ring which is firing . this fact is very important for our analysis . [ one ] assume @xmath121 satisfy ( [ teq]),([req ] ) with conditions ( [ bnp]),([bnr]),([fix ] ) . then @xmath122 @xmath123 ( [ one2 ] ) follows from ( [ one1 ] ) by ( [ bnp ] ) . to prove ( [ one1 ] ) , we first note that certainly @xmath124 , since @xmath125 and ( [ teq ] ) imply @xmath78 is constant , contradicting ( [ bnp ] ) . we note the key fact that , by ( [ teq]),([hpi ] ) and ( [ wz ] ) , @xmath126 by lemma [ cut ] , ( [ kf ] ) implies that the equation @xmath127 has at most one solution for each @xmath2 . in particular , since @xmath128 , @xmath129 we have @xmath130 for @xmath131 , and by continuity of @xmath78 this implies ( [ one1 ] ) . let us note that if we knew that for rotating waves the function @xmath78 must be monotone , then lemma [ one ] would follow immediately from ( [ fix ] ) . is it true in general that rotating wave solutions are monotone ( for m>0 ) ? [ lampos ] assume @xmath121 satisfy ( [ teq]),([req ] ) with conditions ( [ bnp]),([bnr]),([fix ] ) . then @xmath132 . in other words , @xmath133 for all rotating waves with @xmath42 , so the waves rotate clockwise . of course in the symmetric case @xmath134 the waves will rotate counter - clockwise . by ( [ fix ] ) and ( [ kf ] ) we have @xmath135 . we have already noted that @xmath124 . if @xmath136 were negative , then @xmath82 would be decreasing near @xmath137 , so for small @xmath138 we would have @xmath139 , contradicting ( [ one1 ] ) . our next step is to solve ( [ req]),([bnr ] ) for @xmath79 , in terms of @xmath78 . we will use the following important consequence of lemma [ one ] : [ ret ] @xmath140 by lemma [ one ] we have @xmath141 so we will show that @xmath142 let @xmath143 be a test function . using lemma [ one ] again we have @xmath144 where @xmath145 is arbitrary . in particular , since @xmath146 , we may choose @xmath145 sufficiently small so that @xmath147 for @xmath148 , so that we can make a change of variables @xmath149 , obtaining @xmath150 this proves ( [ jon ] ) , completing the proof of the lemma . by lemma [ ret ] we can rewrite equation ( [ req ] ) on the interval @xmath151 as @xmath152 the solution of which is given by @xmath153 where @xmath154 is the heaviside function : @xmath155 for @xmath156 , @xmath157 for @xmath138 . substituting @xmath158 into ( [ gs ] ) and using ( [ bnr ] ) , we obtain an equation for @xmath159 whose solution is @xmath160 and substituting this back into ( [ gs ] ) , we obtain that the solution of ( [ req]),([bnr ] ) which we denote by @xmath161 in order to emphasize the dependence on the parameter @xmath136 , is given on the interval @xmath151 by @xmath162 \;\;\;\ ; 0<|z|<2\pi.\ ] ] we note that , for general @xmath163 , @xmath161 is given as the @xmath164-periodic extension of the function defined by ( [ fr ] ) from @xmath165 $ ] to the whole real line . the following result , which can be computed from ( [ fr ] ) , will be needed later @xmath166 where @xmath167 we note that @xmath168 a fact that considerably simplifies the formulas in the case @xmath24 . the rotating waves correspond to solutions @xmath169 of the equation @xmath170 with @xmath78 satisfying ( [ fix ] ) and @xmath171 to simplify notation , we define @xmath172 so that ( [ twe0 ] ) is rewritten as @xmath173 we note that ( [ twe ] ) is a nonautonomous differential equation for @xmath78 , and since the nonlinearities are bounded and lipschitzian , the initial value problem ( [ twe]),([fix ] ) has a unique solution , which we denote by @xmath174 . rotating waves thus correspond to solutions @xmath132 of the equation @xmath175 rewriting ( [ twe ] ) and ( [ ip ] ) we have @xmath176 @xmath177 and defining @xmath178 we obtain that rotating waves correspond to solutions @xmath132 of the equation @xmath179 assuming that the coupling is @xmath180 we shall be able to solve for the rotating waves explicitly . in this case we have , from ( [ defrl]),([mv ] ) @xmath181 so that ( [ tww ] ) reduces to @xmath182 the fact that ( [ tww0 ] ) is an autonomous equation is what makes the treatment of the case @xmath11 constant much simpler . indeed , assume that ( [ tt ] ) holds , so that @xmath183 then we have , using ( [ tww0 ] ) , making a change of variables @xmath184 , and using ( [ sta ] ) @xmath185 substituting the explicit expressions for @xmath186 and @xmath187 from ( [ spec ] ) , and using the formula @xmath188 ( [ lam ] ) becomes @xmath189 so that rotating waves correspond to solutions of ( [ eqc ] ) , with their velocities given by @xmath190 . we can rewrite ( [ eqc ] ) as @xmath191 where @xmath192 in the following lemma we collect some properties of the functions @xmath193 , which are obtained by elementary calculus : [ prf ] ( i ) when @xmath5 , @xmath194 is positive and convex on @xmath195 , and @xmath196 @xmath197 \(ii ) when @xmath198 , @xmath194 is decreasing on @xmath195 , and ( [ lz ] ) holds . if @xmath7 it has a zero at @xmath199 , if @xmath200 it is positive on @xmath195 and @xmath201 . from lemma [ prf ] we conclude that when @xmath5 ( [ eqc1 ] ) has exactly two solutions if @xmath202 , where @xmath203 which we will denote by @xmath204 no solution if @xmath205 , and a unique solution when @xmath206 . when @xmath198 , part ( ii ) of lemma [ prf ] implies that ( [ eqc1 ] ) has a unique solution for any @xmath22 , which we denote by @xmath207 . an elementary asymptotic analysis of the equation ( [ eqc1 ] ) yields [ asa ] ( i ) when @xmath5 we have the following asymptotics as @xmath208 @xmath209 for @xmath210 , in case @xmath23 we have @xmath211 while in case @xmath24 we have @xmath212 \(ii ) for @xmath198 , the asymptotics of @xmath207 as @xmath213 are the same as those of @xmath210 , given in ( [ asl2 ] ) for @xmath23 and ( [ asl3 ] ) for @xmath24 . we thus obtain [ unij ] when @xmath180 : \(i ) in the excitable case @xmath214 : \(i ) if @xmath215 there exist two rotating waves with velocities given by @xmath216 and we have , for the slow wave @xmath217 for the fast wave when @xmath23 : @xmath218 while for the fast wave when @xmath24 @xmath219 ( ii ) if @xmath206 there exists a unique rotating wave with velocity @xmath220 ( iii ) if @xmath205 there exist no rotating waves . \(ii ) when @xmath198 , there exists a unique rotating wave for any @xmath22 , whose velocity is given by @xmath221 and for large @xmath27 it has the same asymptotics as in ( [ asv2]),([asv22 ] ) in the cases @xmath23 , @xmath24 , respectively . in the excitable case we thus have two rotating waves born at a supercritical saddle - node bifurcation as the coupling strength @xmath27 crosses @xmath222 . we now note that in the special case @xmath24 ( the model introduced in @xcite ) we can obtain more explicit expressions . using ( [ rho0 ] ) we have @xmath223 the minimum in ( [ defom ] ) can now be computed explicitly , and we obtain , when @xmath5 , @xmath224 we can also solve ( [ eqc1 ] ) explicitly , and obtain the velocities of the rotating waves . when @xmath5 , @xmath225 @xmath226 when @xmath198 , for all @xmath22 @xmath227 figures [ syn1],[syn2 ] show the wave - velocity vs. coupling strength diagrams for the rotating waves when @xmath228 , @xmath24 , in an excitable ( @xmath229 ) and an oscillatory ( @xmath230 ) case . in figures [ syn3],[syn4 ] we change @xmath231 to @xmath232 . ) vs. coupling strength ( @xmath27 ) for the case @xmath180 , @xmath24 , @xmath229.,width=264,height=264 ] ) vs. coupling strength ( @xmath27 ) for the case @xmath180 , @xmath24 , @xmath230.,width=264,height=264 ] ) vs. coupling strength ( @xmath27 ) for the case @xmath180 , @xmath232 , @xmath229.,width=264,height=264 ] ) vs. coupling strength ( @xmath27 ) for the case @xmath180 , @xmath232 , @xmath230.,width=264,height=264 ] we now return to the case when @xmath11 is a general continuous positive @xmath164-periodic function , and prove that several of the results about rotating waves obtained above for the special case @xmath180 remain valid , though the proofs are necessarily less direct . for @xmath24 , @xmath161 becomes singular as @xmath239 , so we need a more refined argument . for @xmath240 we have @xmath241 uniformly in @xmath235 $ ] , hence @xmath242 where @xmath243 the standard theorems on dependence of solutions of initial - value problems on parameters hence imply @xmath244 uniformly in @xmath235 $ ] , where @xmath245 satisfies @xmath246 and @xmath247 . since , by ( [ wz ] ) , the constant function @xmath248 is a solution to this initial - value problem , the uniqueness theorem for initial - value problems implies that @xmath249 . in the case of constant @xmath11 the result can be proven by direct computation , so we now assume @xmath11 is not a constant function . we will show that when @xmath253 we have @xmath254,\ ] ] @xmath255.\ ] ] together with ( [ aps ] ) , these imply the result of our lemma . to prove our claim we note that , using ( [ tww]),([hz]),([wnn ] ) , part ( ii ) of lemma [ ulb ] ( which is why we need the assumption that @xmath11 is non - constant ) and ( [ cl ] ) @xmath256 we now show that ( [ kee ] ) implies ( [ cons1 ] ) . if ( [ cons1 ] ) fails to hold , then we set @xmath257 \;|\ ; \phi_{\lambda}(z)= 2\pi\}.\ ] ] this number is well - defined by continuity and by the fact that @xmath258 , which implies also that @xmath259 . by ( [ kee ] ) we have @xmath260 , but this implies that @xmath261 is decreasing in a neighborhood of @xmath69 , and in particular that there exist @xmath262 satisfying @xmath263 but this contradicts the definition of @xmath69 , and this contradiction proves ( [ cons1 ] ) . similarly , assuming ( [ cons2 ] ) does not hold and defining @xmath264 \;|\ ; \phi_{\lambda}(z)=0\},\ ] ] we conclude that @xmath265 and @xmath266 , so that @xmath261 is decreasing in a neighborhood of @xmath267 , and this implies a contradiction to the definition of @xmath267 and proves that ( [ cons2 ] ) holds . this concludes the proof of the lemma . let us note that since @xmath272 implies that ( [ tt ] ) does nt hold , and since @xmath190 , we can reformulate the previous lemma as a lower bound for the velocities of rotating waves in the excitable case . we shall show that if @xmath276 then @xmath277 for all @xmath132 , and thus that equation ( [ tt ] ) can not hold . by ( [ aps ] ) , ( [ sm ] ) is equivalent to @xmath278 we note that , by lemma [ ell ] , we already have ( [ sm ] ) when ( [ cl ] ) holds , hence we may assume @xmath279 we define @xmath280 and we note that ( [ aaa ] ) is equivalent to the statement that @xmath281 using ( [ tww ] ) and lemma [ ulb ] we have @xmath282 \nonumber\\&\leq & \lambda \big[1-\cos(\phi_{\lambda}(x))+ \big(\beta+\frac{g}{2\lambda}\rho_c(\lambda)\max_{x\in{\mathbb{r}}}{j(x)}\big ) ( 1+\cos(\phi_{\lambda}(z)))\big]\nonumber\\&= & \lambda [ ( \mu + 1)+(\mu-1)\cos(\phi_{\lambda}(z))].\end{aligned}\ ] ] which implies ( note that the integral below is well - defined because of ( [ mgz ] ) ) @xmath283 making the change of variables @xmath184 we obtain @xmath284 if we assume , by way of contradiction , that ( [ din0 ] ) does not hold , _ i.e. _ that @xmath285 , then , using ( [ form ] ) , @xmath286 so together with ( [ ha0 ] ) we obtain @xmath287 which is equivalent to @xmath288 which contradicts @xmath289 . this contradiction proves ( [ din0 ] ) , concluding the proof of the theorem . [ two ] in the excitable case @xmath5 , if there exists some @xmath290 with @xmath291 then there exist at least two solutions @xmath292 of ( [ tt ] ) with @xmath293 , hence two rotating waves , with velocities satisfying @xmath294 by lemma [ psiprop ] , we can choose @xmath295 so that @xmath296 . by lemma [ ell0 ] , if we fix @xmath297 then @xmath298 for all @xmath299 , and in particular it follows that @xmath300 . we thus have @xmath301 with @xmath302 thus by the intermediate value theorem , the equation ( [ tt ] ) has a solution @xmath303 and a solution @xmath304 , corresponding to two rotating waves . by ( [ aps ] ) , our claim is equivalent to @xmath308 we define @xmath309 and we note that ( [ kein ] ) is equivalent to @xmath310 using ( [ tww ] ) and lemma [ ulb ] we have @xmath311 \nonumber\\&\geq & \lambda\big[1-\cos(\phi_{\lambda}(x))+ \big(\beta+\frac{g}{2\lambda}\rho_c(\lambda)\min_{x\in{\mathbb{r}}}{j(x)}\big ) ( 1+\cos(\phi_{\lambda}(z)))\big]\nonumber\\&= & \lambda [ ( \eta + 1)+(\eta-1)\cos(\phi_{\lambda}(z))],\end{aligned}\ ] ] which implies @xmath312 making the change of variables @xmath184 , we obtain @xmath313 if we assume , by way of contradiction , that ( [ din ] ) does not hold , _ i.e. _ that @xmath314 , then , using ( [ form ] ) , @xmath315 so together with ( [ ha ] ) we obtain @xmath316 this contradicts ( [ eg2 ] ) , and this contradiction implies that ( [ din ] ) holds , completing our proof . [ ts ] in the excitable case @xmath5 , let @xmath317 where @xmath222 is defined by ( [ defom ] ) . then when @xmath318 , there exist at least two rotating waves . in fact , we have a ` slow ' wave with velocity @xmath319 bounded from above by @xmath320 and a ` fast wave ' with velocity @xmath321 bounded from below by @xmath322 where @xmath323 are the functions defined by ( [ vdef ] ) . to prove ( [ sw]),([fw ] ) , we note that , assuming @xmath318 , the range of values of @xmath328 for which ( [ kein ] ) holds is the interval @xmath329 where the functions @xmath330 are defined in section [ constant ] . thus , applying lemma [ two ] with @xmath331 where @xmath145 is arbitrarily small , we obtain the existence of a solution @xmath332 of ( [ tt ] ) with @xmath333 . since @xmath145 is arbitrary , we have a solution @xmath136 of ( [ tt ] ) with @xmath334 , hence a rotating wave with velocity @xmath319 satisfying ( [ sw ] ) . similarly applying lemma [ two ] with @xmath335 , we obtain the existence of a wave with velocity @xmath321 satisfying ( [ fw ] ) . theorems [ noex ] and [ ts ] show that several of the qualitative features that we saw explicitly in the case of uniform coupling ( section [ constant ] ) remain valid in the general case . it is natural to ask whether more can be said , _ e.g. _ , whether the following conjecture , or some weakened form of it , is true : for any @xmath11 , there exists a value @xmath336 such that : [ osct ] [ osc ] if @xmath198 , there exists a rotating wave solution for any value of @xmath22 , with velocity @xmath39 bounded from below by @xmath340 where @xmath39 is the function defined by ( [ vdef1 ] ) , and the asymptotic formulas ( [ fwe0]),([fwe ] ) hold with @xmath321 replaced by @xmath39 . if @xmath7 , then for any @xmath22 the equation @xmath341 has the unique solution @xmath342 hence , any @xmath343 satisfies ( [ kein ] ) , so that by lemma [ glar ] @xmath344 . on the other hand for @xmath136 sufficiently small we have , by lemma [ psiprop ] , @xmath272 . hence there exists a solution @xmath345 of ( [ tt ] ) . since @xmath346 is arbitrary , we conclude that there exists a solution @xmath347 of ( [ tt ] ) . hence a rotating wave with velocity satisfying ( [ fw1 ] ) . investigate the question of stability of the rotating waves , _ i.e. _ , do arbitrary solutions of ( [ ge ] ) , ( [ syn ] ) approach one of the rotating waves in large time ? we conjecture that , at least under some restrictions on @xmath11 , the rotating wave is stable in the case @xmath7 , while in the case @xmath5 the fast rotating wave is stable and the slow one is unstable . izhikevich , _ class 1 neural excitability , conventional synapses , weakly connected networks , and mathematical foundations of pulse - coupled models _ , ieee trans . neural networks * 10 * ( 1999 ) , 499 - 507 .

we study rotating waves in the theta model for a ring of synaptically - interacting neurons . we prove that when the neurons are oscillatory , at least one rotating wave always exists . in the case of excitable neurons , we prove that no travelling waves exist when the synaptic coupling is weak , and at least two rotating waves , a ` fast ' one and a ` slow ' one , exist when the synaptic coupling is sufficiently strong . we derive explicit upper and lower bounds for the ` critical ' coupling strength as well as for the wave velocities . we also study the special case of uniform coupling , for which complete analytical results on the rotating waves can be achieved .

oscillatory transitions among neutrinos of different flavors have emerged as a major topic of particle physics @xcite . an accelerator experiment using muon antineutrinos from @xmath16 decays at rest over an effective baseline of @xmath17 m / mev , lsnd at los alamos , has reported a positive signal in the channel @xmath18 with a probability of @xmath19 @xcite . a consistent , albeit less significant , signal was also observed in the cp - conjugate channel @xmath5 using muon neutrinos from @xmath20 decay in flight @xcite . when combined with the upper limits imposed by other accelerator and reactor experiments @xcite , the lsnd data suggest that the @xmath18 oscillation is driven by a mass difference squared , @xmath21 , between some 0.3 and 2.3 ev@xmath7 . the transitions @xmath18 and @xmath5 in this region of @xmath21 will soon be explored with higher sensitivity by the boone experiment at fermilab @xcite . qualitatively , an analogy with the quark sector suggests that ( i ) the @xmath22 should pick the largest contribution of the heaviest mass state @xmath23 , and that ( ii ) the mixings between the neighboring neutrino flavors ( _ i.e. _ , @xmath24 and @xmath25 ) should be the strongest . therefore , a mass difference as large as @xmath6 ev@xmath7 should primarily manifest itself in the @xmath4 channel . the existing upper limits on the effective mixing @xmath26 for the transition @xmath4 @xcite largely come from experiments with small effective baselines of @xmath27 km / gev , and therefore are much less compelling for @xmath6 ev@xmath7 than for larger values of the mass difference . at neutrino energies well below the @xmath0 threshold ( @xmath28 gev ) , the transition @xmath4 in the required region @xmath29 km / gev will be indirectly probed by boone through @xmath2 disappearance @xcite . however , a convincing @xmath4 signal can only be demonstrated by detecting the cc collisions of the @xmath22 in appearance mode . for this , an appropriate environment is offered by the `` medium - baseline '' location on mount jura @xcite in the @xmath2 beam of cern - sps ( @xmath30 km and @xmath31 gev by flux ) . the cc collisions of the @xmath22 may be identified either topologically @xcite or by detecting the detached vertex of the produced @xmath0 in a hybrid - emulsion spectrometer @xcite . the latter option is considered in this paper . for the `` original '' transition @xmath5 to be observed and investigated in the same experiment , that will have very different systematics compared to either lsnd @xcite and boone @xcite , secondary electrons must be efficiently detected and momentum - analyzed . the atmospheric evidence for neutrino oscillations driven by a mass difference of 10@xmath3210@xmath10 ev@xmath7 @xcite will be initially probed by the long - baseline experiments operating in the neutrino beams of proton accelerators @xcite . large angular divergences of these beams dictate that , for the sake of statistics , fine instrumentation of the detector and/or the magnetic analysis be compromised for a multi - kiloton fiducial mass . therefore , unraveling the pattern of oscillations in this mass region will probably require a much more intense beam generated by a muon storage ring with a straight section that points towards the detector @xcite . assuming a @xmath33 ring , the beam fully consists of muon neutrinos and electron antineutrinos whose energy spectra and angular spreads are precisely known . that the original beam contains neither @xmath34 nor @xmath35 allows to identify the parent neutrino by the sign of the lepton produced by the oscillated neutrino : negative and positive leptons are unambiguous signatures of the @xmath2 and @xmath36 parents , respectively . therefore , the transitions @xmath5 , @xmath4 , @xmath37 , and @xmath38 can be tagged independently . the unique properties of this neutrino beam may warrant a finely instrumented detector of relatively small mass , that would be sensitive to a broad range of neutrino transitions driven by a mass difference of 10@xmath3210@xmath10 ev@xmath7 . for this , several conditions are obligatory . selecting the @xmath22 and @xmath39 collisions by the detached vertex of the @xmath0 is only possible in a hybrid - emulsion apparatus . in order to suppress the background to @xmath0 decays arising from anticharm production in cc collisions of the @xmath36 , the detector should identify the electrons as reliably as the muons . for the transitions @xmath4and @xmath38 to be reliably discriminated , all charged secondaries ( and not just muons ) should be sign - selected and momentum - analyzed in the detector . the latter will also allow extra kinematic handles for further suppressing the background to @xmath0 decays ( such as @xmath40 of a decay particle with respect to @xmath0 direction and transverse disbalance of the event as a whole with respect to incident neutrino ) . in this paper , we propose a conceptual scheme of a hybrid - emulsion spectrometer for detecting and identifying the neutrinos of different flavors by their cc collisions , as required for probing various channels of neutrino oscillations . the design emphasizes detection of @xmath0 leptons by detached vertices , reliable identification and sign - selection of electrons , good spectrometry for all charged secondaries , and reconstruction of secondary photons and @xmath41 mesons . we also estimate the performance of the proposed apparatus in a medium - baseline experiment using the neutrino beam of the proton machine cern - sps and in a long - baseline experiment in the neutrino beam of a muon storage ring . building on the ideas put forth by a. ereditato , k. niwa , and p. strolin @xcite , we implement the principle of `` emulsion cloud chamber '' : layers of thin emulsion are only used as a tracker for events occurring in passive material . but in contrast with @xcite , we aim at constructing a distributed target with low density and large radiation length @xcite , so that either muons , hadrons , and electrons can be momentum - analyzed by curvature inside the target itself . accordingly , the target should largely consist of low - z material like aluminum or carbon in the form of carbon - fiber composite . apart from narrow gaps instrumented with drift chambers that provide an electronic `` blueprint '' of the event , the target is built as a compact homogeneous volume in ambient magnetic field . owing to relatively weak multiple scattering in low - z material , the successive layers of thin emulsion may act as an `` emulsion spectrometer '' for analyzing the momenta of charged secondaries by curvature in magnetic field . in untangling the topologies of neutrino events , the detector will operate very much like a bubble chamber . the proposed apparatus is therefore referred to as the emulsion bubble chamber , or ebc . specified below is a tentative structure of the distributed target that has been assumed in our simulations of detector response . ( the actual design , including the choice of low - z material , must of course be carefully optimized for a particular experiment ) . the fine structure of the target is depicted in fig . [ structure ] . the 6-mm - thick basic element of the structure is formed by a 1-mm plate of passive material , by an emulsion plastic emulsion sheet ( referred to as the es ) with total thickness of 0.2 mm , and by a relatively large drift space of 4.8 mm in which @xmath0 decays are selected . in our tentative design , the passive plate is largely carbon ( 960 @xmath42 m ) , but also includes a thin layer of denser substance ( 40 @xmath42 m of copper ) in order to boost the geometric acceptance . the es is a 100-@xmath42 m sheet of transparent plastic coated on both sides with 50-@xmath42 m layers of emulsion . the medium formed by successive elements has mean density of @xmath43 g/@xmath44 and effective radiation length of @xmath45 cm . ( these values of @xmath46 and @xmath47 are very similar to those of a bubble chamber with neon - hydrogen filling @xcite . ) note that the element does not feature the second es downstream of the gap , as originally foreseen in @xcite : the idea is to detect the kink or the trident using track segments in the ess of two successive elements , as allowed by relatively weak coulomb scattering in low - z material of the intervening passive plate . on the technical side , the air gap can be created by thin and rigid `` bristles '' on the upstream face of the carbon - composite plate that has been manufactured in `` brush - like '' form . ( this is only possible because we have just one es per element . ) the positions of the thin bristles will be tabulated , and secondary vertices that match these positions will be dropped . another technical option for the drift space is 5-mm - thick paper honeycomb . for a @xmath0 that emerged from the passive plate and suffered a one - prong or a three - prong decay in the gap , the decay vertex can be reconstructed from the track segments in two successive ess . as soon as the fitted secondary vertex lies within the intervening passive plate , the candidate event must be dropped because of the high background from reinteractions . in principle , a @xmath0 that decayed before reaching the gap can be detected by impact parameter , but this possibility is not considered here . by adding a thin layer of denser material ( copper ) downstream of low - z material ( carbon ) , we slightly compromise the radiation length for geometric acceptance : thereby , the fraction of @xmath0 leptons that reach the air gap is effectively increased . that the proportion of copper events is boosted by the geometric effect is illustrated by fig . [ vertices ] . here and below , the spectrum of incident @xmath0 neutrinos is assumed to be proportional to the spectrum of muon neutrinos from cern - sps @xcite . for the tracks to be found in emulsion at scanning stage , planes of electronic detectors must be inserted in the continuous structure formed by the basic elements . these detectors should allow a crude on - line reconstruction of the event as a whole , and therefore they must have sufficient spatial resolution , provide angular information , and be spaced by less than one radiation length @xmath47 . in our tentative design , a stack of 30 elements forms a `` module '' with total thickness of @xmath48 , and 4-cm - wide gaps between adjacent modules are instrumented by multisampling drift chambers . in a medium- or long - baseline experiment where the occupancy of ess will be relatively low , the accuracy of `` electronic '' reconstruction should be sufficient for unambiguously finding an energetic track in emulsion . therefore , one can scan back along a stiff track , starting from the downstream es of the module in which the collision occurred . as soon as the layer of origin is reached , a few successive ess of the nearby elements must be fully scanned over relatively small areas towards finding the stubs of all other tracks associated with the primary vertex . to refine the alignment of ess , a sufficient number of stiff muon tracks must be fully reconstructed in emulsion . this first stage of emulsion scanning will yield a relatively small number of events featuring decay signatures ( either a kink or a trident in the air gap just downstream of the primary vertex ) . the second stage will be to scan down all tracks emerging from the primary vertex and to find and measure the conversions of secondary photons in the distributed target . this will allow to analyze the momenta of the secondaries in the `` emulsion spectrometer '' and to identify electrons by change of curvature and by emission of brems . we foresee that the scanning of emulsion will only start after the full period of detector exposure . the response of the ebc detector is simulated using geant . in fitting a track , we assume that each tracker traversed ( either a es or a drift chamber ) provides two spatial points with resolutions of 2 and 150 @xmath42 m , respectively . because of multiple scattering , the fit is but marginally sensitive to increasing the spatial error in the es from 2 up to 8 @xmath42 m . a uniform magnetic field of 0.7 tesla , that is normal with respect to beam direction , is assumed throughout . ( using a stronger field would boost the performance of the spectrometer , but may be impracticable for a big air - core magnet that will house a multiton ebc . ) when found in emulsion , the track of an electron can be identified by a variation of curvature due to energy losses in successive layers of the target . these losses must be accounted for in estimating the momentum by curvature in magnetic field . for this , we use a simple algorithm that should be treated as preliminary and is not fully realistic . namely , we fit a restricted segment of the @xmath49 track over which the actual loss of energy does not exceed 30% . additionally , this segment is required to cross no less than five ess . we compute an `` ideal trajectory '' for a given value of electron momentum @xmath50 , to which the observed trajectory is then fitted , using geant . in doing so , we switch off multiple scattering and radiation losses , but instead reduce the momentum `` by hand '' in each layer of the target by the same amount @xmath51 that is treated as an empirical parameter . the value of @xmath51 is selected so as to obtain ( i ) an unbiased estimate of electron momentum ( @xmath52 ) and ( ii ) a reasonable value of @xmath53 . for definiteness , we use a monte - carlo template of electrons with @xmath54 gev originating from simulated decays @xmath1 ( see further ) . for these electrons , the mean length of the selected track segment is close to 30 cm ( see fig . [ length ] ) , which proves to be sufficient for analyzing the momentum in the `` emulsion spectrometer '' formed by successive ess of the distributed target . plotted in fig . [ aluminum ] is the ratio between the fitted and true momenta , @xmath55 , for the properly fitted electrons ( @xmath56 ) . the same ratio is then separately shown for two regions of electron momentum : @xmath57 gev and @xmath58 gev . we estimate that of all electrons with @xmath54 gev , nearly 90% can be reliably detected , identified , and sign - selected ( that is , reconstructed with @xmath56 and @xmath59 ) . on average , the @xmath49 momentum is measured to a precision of some 11% . treating the length of the track segment as a free parameter of the fit , which is a more realistic approach that is not attempted in this paper , will further improve the momentum resolution and boost the fraction of sign - selected electrons . good spectrometry for electrons will allow ebc to detect and discriminate the cc collisions of electron neutrinos and antineutrinos and , at the same time , to select electronic decays of the @xmath0 almost as efficiently as the muonic ones . a comparison between the two leptonic modes will provide an important handle on self - consistency of any @xmath0 signal . the spectrometry of muons and charged pions is much less affected by energy loss in matter , but for pions the momentum resolution is slightly degraded by hadronic reinteractions . for the pions with @xmath60 gev originating from the simulated decays @xmath61 , fig . [ momres ] shows the ratio between the fitted and true momenta , @xmath62 . the mean uncertainty on pion momentum is seen to be close to 7% . neutral pions can be reconstructed from photon conversions in the distributed target . ( here , we assume that the potential length in the detector is much larger than @xmath45 cm . ) in estimating the energy of a photon that has converted in the target , we count only those conversion electrons that have fired at least one drift chamber and , therefore , can be found in emulsion and then momentum - analyzed by curvature . we also assume that the primary vertex has already been fitted , so that the direction of the photon is precisely known . for illustration , we consider the @xmath41 mesons originating from the decay @xmath63 . the measured invariant mass of the two detected photons from _ gg , as plotted in fig . [ mgg ] , shows a distinct @xmath41 signal . the actual size of the mass window for selecting the @xmath41 candidates will be dictated by the level of combinatorial background in a particular analysis ; for purely illustrative purposes , we assume a mass window of @xmath64 mev . thus estimated detection efficiency for @xmath41 mesons is close to 0.26 . for the @xmath0 leptons emitted in either the deep - inelastic and quasielastic @xmath65 collisions , we generate the two leptonic decays , @xmath66 and @xmath1 , and three semileptonic ( quasi-)two - body decays : @xmath61 , @xmath63 that is mediated by the resonance _ 2pi , and 1dec that is mediated by the resonance 1_3pi . the threshold effect for @xmath0 production in neutrino nucleon collisions and polarization of the @xmath0 , that affects the angular distribution of decay products in the @xmath0 frame , are accounted for . further details on our @xmath0 generator can be found in @xcite . for definiteness , we again assume that @xmath0 neutrinos incident on ebc have the same energy spectrum as muon neutrinos from cern - sps @xcite ( @xmath31 gev by flux ) , and that the distributed target is magnetized by a uniform field of 0.7 tesla that is perpendicular to beam direction . the decay products of the @xmath0 are propagated through the target and then reconstructed from hits in emulsion , as explained in the previous section . in estimating the detection efficiencies for different decay channels of the @xmath0 , we adopt the ( quasi - realistic ) selection criteria that are listed below . * the @xmath0 must have decayed in the drift gap . this is necessary for reconstructing the detached vertex from track segments in the upstream and downstream ess . * the momentum of either charged daughter must exceed 1 gev , and its emission angle must lie within 400 mrad of beam direction . these selections are suggested by the fact that soft and broad - angle tracks are poorly reconstructed in emulsion . * for a one - prong decay to a charged daughter @xmath67 ( either a muon , electron , or pion ) , the kink angle must be suffiently large : @xmath68 mrad . this lower cut reflects the experimental uncertainty on the kink angle that is close to 5 mrad . * all charged daughters of the @xmath0 must be momentum - analyzed by curvature and reliably sign - selected ( that is , reconstructed in the detector with @xmath56 and @xmath69 ) . this will fix the sign of the @xmath0 and , on the other hand , will allow kinematic handles for rejecting the decays of charmed and strange particles . * in a one - prong decay , @xmath40 of the charged daughter with respect to @xmath0 direction must exceed 250 mev . this is aimed at rejecting the decays of strange particles . for those @xmath0 decay channels that have actually been simulated in the detector , the estimated detection efficiencies ( or acceptances ) are listed in table [ accept ] . ( note that here we do not require the @xmath41 from @xmath63 to be reconstructed . ) for these decay channels of the @xmath0 , the acceptance - weighted branching fractions add up to some 0.28 . approximately accounting for the other one - prong and three - prong channels , we estimate that nearly 32% of all @xmath0 leptons produced in passive material will be detected in ebc by visible kinks or tridents . .the detection efficiencies , or acceptances , for those decay channels of the @xmath0 that have been simulated in the detector . [ cols="^,^,^,^,^ " , ] the ( quasi-)two - body decays @xmath61 , @xmath70 , and 1dec may provide an extra kinematic handle for discriminating the @xmath71 against the background of anticharm . in a decay @xmath72 , `` transverse mass '' is defined as @xmath73 , where @xmath74 and @xmath40 are the @xmath75 mass and transverse momentum with respect to @xmath0 direction . the two - body kinematics dictate that the unsmeared @xmath76 distribution should reveal a very distinctive peak just below @xmath77 , see the upper plots of fig . the @xmath76 technique for identifying massive parents by two - body decays in emulsion was proven by discriminating the relatively rare decay @xmath78 against a heavy background from other decays of charm @xcite . observing the high-@xmath76 peak in a detector requires good spectrometry of charged pions and , for the decay @xmath63 in particular , good reconstruction of @xmath41 mesons . the bottom plots of fig . [ m_t ] show the smeared @xmath76 distributions for the reconstructed decays @xmath61 , @xmath63 , and 1dec ( the @xmath41 for @xmath63 has been selected in the mass window @xmath64 mev , see fig . [ mgg ] ) . the jacobian cusps of the original @xmath76 distributions largely survive the apparatus smearings of ebc and , therefore , will provide distinctive signatures of the @xmath0 . neutrino oscillations driven by @xmath6 ev @xcite can be efficiently probed at a location on mount jura near cern , that is irradiated by the existing wide - band @xmath2 beam of the cern - sps accelerator ( @xmath31 gev by flux ) over a `` medium '' baseline of 17 km @xcite . at this location , an ebc detector can be deployed in the existing magnet of the nomad detector @xcite that delivers a magnetic field of up to 0.7 tesla . in the large internal volume of this magnet , @xmath79 m@xmath80 , one might deploy a distributed target with a volume of @xmath81 m@xmath80 , that will consist of 30 modules ( see section 2 ) . the target has a total thickness of 10 radiation lengths , and contains nearly 20 tons of passive material and 3 tons of standard emulsion ( or a lesser amount of diluted emulsion , that may be warranted by a relatively low occupancy at a medium - baseline location ) . the magnetic volume will also house a few drift chambers downstream of the target . as photon conversions and electrons will be efficiently detected in the distributed target , no extra electromagnetic calorimeter is foreseen . the design of the muon system is not discussed in this paper . at the jura location , the rate of @xmath2-induced cc collisions has been estimated @xcite as 843 events per ton of target per @xmath82 protons delivered by cern - sps . assuming @xmath83 delivered protons which corresponds to 34 years of operation , the proposed detector will collect nearly @xmath84 cc events . even if the probability of the @xmath4 transition is as small as 0.3% @xcite , we will detect some 80 @xmath0 events with a negligibly small background from the decays of strange and anticharm particles . ( this prediction assumes a @xmath0 detection efficiency of 32% , as estimated above for the same beam , and takes into account the threshold effect in @xmath0 production . ) alternatively , a zero signal will allow to exclude ( at 90% c.l . ) an area of the parameter plane ( @xmath85 , @xmath86 ) that is depicted in fig . [ exclude ] . as demonstrated above , ebc will very efficiently identify , sign - select , and momentum - analyze prompt electrons from cc collisions of electron neutrinos and antineutrinos . the energy of incident @xmath35 ( @xmath87 ) will be estimated to better than 10% . if at the jura site the transition @xmath5 indeed occurs with a probability of some 0.003 as in lsnd @xcite , in an exposure of @xmath83 protons on target we will detect and reconstruct some 460 @xmath88 events due to oscillated neutrinos against a background of nearly 860 cc events due to the original @xmath35 component of the beam @xcite . as the original @xmath35 component of the beam is substantially harder than the @xmath2 component , a @xmath5 signal will effectively reduce the mean energy of @xmath88 events . a very intense and collimated beam of a muon storage ring @xcite will allow long - baseline experiments with relatively small but finely instrumented detectors . for purely illustrative purposes , we assume an ebc detector as sketched above ( 20 tons of passive material plus 3 tons of standard emulsion ) that is irradiated by a @xmath33 ring with parameters as foreseen in @xcite ( that is , @xmath89 injected muons per year of operation and a straight section that amounts to 25% of the ring circumference ) . we also assume a `` nominal '' distance of 732 km ( either from cern to gran sasso or from fermilab to soudan ) that results in effective baselines of @xmath90 and 10 km / gev for the stored - muon energies of @xmath91 and 100 gev , respectively . then , given unpolarized muons with @xmath91 ( 100 ) gev in the ring , in the absence of oscillations the detector will annually collect some @xmath92 ( @xmath93 ) and @xmath94 ( @xmath95 ) cc collisions of muon neutrinos and electron antineutrinos , respectively . note that the expected event rate for @xmath96 gev is as high as in the cern - sps beam at jura location . taken together , the data on atmospheric @xcite and reactor @xcite neutrinos favor a @xmath4 transition with almost maximal mixing and with a mass difference in the range @xmath9710@xmath10 ev@xmath7 @xcite . for definiteness assuming @xmath98 and @xmath99 ev@xmath7 , we estimate that in three years of operation a 20-ton ebc will detect a signal of some 50 ( 170 ) @xmath71 leptons for @xmath91 ( 100 ) gev with a negligible background of @xmath36-produced anticharm . ( note that the @xmath100-dependence of the @xmath0 signal largely reflects the increase of @xmath101 with neutrino energy : in the region @xmath21(ev@xmath7 ) @xmath102 ( gev / km ) , the flux of oscillated neutrinos is virtually independent of @xmath100 . ) alternatively , a zero @xmath71 signal will effectively exclude a large area of parameter space for the transition @xmath4 , see fig . [ exclude ] . the transition @xmath38 will be simultaneously detected with a comparable sensitivity , and the two transitions will be reliably discriminated by the charge of the @xmath0 . using the detached - vertex information , the transitions @xmath5 and @xmath37 will be separated from the transitions @xmath4 and @xmath38 involving the leptonic decays @xmath1 and @xmath103 , respectively . we may conclude that , in this experimental environment , even a relatively small ebc detector will efficiently probe the entire pattern of neutrino oscillations in the region @xmath910@xmath10 ev@xmath7 . a conceptual detector scheme is proposed for studying various channels of neutrino oscillations . the hybrid - emulsion spectrometer will detect and discriminate the neutrinos of different flavors by their cc collisions . the design emphasizes detection of @xmath0 leptons by detached vertices , identification and sign - selection of electrons , and spectrometry for all charged particles and photons . a distributed target is formed by layers of low - z material , emulsion - plastic - emulsion sheets , and air gaps in which @xmath0 decays are detected . target modules with mean density of 0.49 g/@xmath44 and radiation length of 52.6 cm , that are similar to those of a bubble chamber with neon - hydrogen filling , are alternated by multisampling drift chambers that provide electronic tracking in real time . the tracks of charged secondaries , including electrons , are momentum - analyzed by curvature in magnetic field using hits in successive thin layers of emulsion and in drift chambers . electrons are identified by change of curvature and by emission of brems . photons are detected and analyzed by conversions in the distribute target , and neutral pions are reconstructed . the @xmath0 leptons are efficiently detected and sign - selected in all major decay channels , including @xmath1 . at a medium - baseline location on mount jura in the existing neutrino beam of the proton machine cern - sps , the detector will be sensitive to either the @xmath4 and @xmath5 transitions in the mass - difference region of @xmath6 ev@xmath7 , as suggested by the results of lsnd . at a long - baseline location in the neutrino beam of a muon storage ring , even a relatively small spectrometer of the proposed type will efficiently probe the entire pattern of neutrino oscillations in the region @xmath910@xmath10 ev@xmath7 that is suggested by the data on atmospheric neutrinos . 99 k. zuber , _ on the physics of massive neutrinos , _ physics reports 305 ( 1998 ) 295 ; + j. brunner , _ experimental results on neutrino masses and lepton mixing , _ preprint cern - ppe/97 - 38 , april 1997 . c. athanassopoulos _ et al . , _ lsnd coll . , phys . 77 ( 1996 ) 3082 . c. athanassopoulos _ et al . , _ lsnd coll . ( 1998 ) 1774 . b. achkar _ et al . , _ bugey coll . b 434 ( 1995 ) 503 . l. borodovsky _ et al . , _ e776 coll . 68 ( 1992 ) 274 . e. church _ et al . , _ boone coll . , _ a proposal for an experiment to measure @xmath5 oscillations and @xmath2 disappearance at the fermilab booster : boone , _ december 7 , 1997 . f. dydak _ et al . , _ cdhs coll . b 134 ( 1984 ) 281 . n. ushida _ et al . , _ e531 coll . 57 ( 1986 ) 2897 . e. eskut et al . , chorus coll . b 424 ( 1998 ) 202 ; preprint cern - ep/98 - 73 , may 1998 ; + p. astier et al . , nomad coll . b 431 ( 1998 ) 219 ; preprint cern - ep/99 - 32 , february 1999 . et al . , _ icarus coll . , _ a search program for explicit neutrino oscillations at long and medium baselines with icarus detectors , _ icarus - tm-97/01 , march 1997 . j. altegoer _ et al . , _ nomad coll . instr . and meth . a 404 ( 1998 ) 96 . y. fukuda _ et al . , _ kamiokande coll . b 335 ( 1994 ) 237 ; s. hatakeyama _ et al . , _ kamiokande coll . 81 ( 1998 ) 2016 ; + y. fukuda _ et al . , _ superkamiokande coll . 81 ( 1998 ) 1562 ; phys . b 433 ( 1998 ) 9 ; phys . lett . b 436 ( 1998 ) 33 ; + m. ambrosio _ et al . , _ macro coll . , phys . b 434 ( 1998 ) 451 . et al . , _ , preprint kek 97 - 266 , january 1998 . e. ables _ et al . , _ minos coll . , fermilab proposal p875 , february 1995 ; minos coll . , _ the minos detector technical design report , _ fermilab report numi - l-337 , october 1998 . s. geer , in _ proc . workshop on physics at the first muon collider and at the front end of a muon collider , _ fermilab , november 1997 , published in aip conf . proc . 435 ( 1998 ) ; phys . rev . d 57 ( 1998 ) 6989 . a. ereditato , k. niwa , and p. strolin , preprint infn / ae-97 - 06 , nagoya dapnu-97 - 07 , june 1997 . et al . , _ e180 coll . b 177 ( 1981 ) 365 ; + k. varvell _ et al . , _ wa59 coll . , z. phys . c 36 ( 1987 ) 1 . a. asratyan _ et al . , _ nucl . instr . and meth . a 427 ( 1999 ) 495 . k. kodama _ et al . , _ e653 coll . , preprint dpnu-96 - 33 , june 1996 . m. apollonio _ et al . , _ chooz coll . b 420 ( 1998 ) 397 . m.c . gonzalez - garcia , h. nunokawa , o.l.g . perez , t. stanev , and j.w.f . valle , phys . d 58 ( 1998 ) 033004 . o. yasuda , phys . rev . d 58 ( 1998 ) 091301 . fogli , e. lisi , a. marrone , and g. scioscia , phys . d 59 ( 1999 ) 033001 . fig . [ structure ] . + schematic of the assumed fine structure of the target , showing the 1-mm - thick carbon copper plates ( 960 + 40 @xmath42 m ) , emulsion plastic emulsion sheets ( 50 + 100 + 50 @xmath42 m ) , and drift space in which @xmath0 decays are selected ( 4800 @xmath42 m ) . [ vertices ] . + the depth of the primary vertex in the carbon copper plate for all @xmath0 events ( a ) and for those events in which the @xmath0 has decayed in the drift space ( b ) . here and in subsequent figures , the energy spectrum of incident @xmath0 neutrinos is assumed to be proportional to the spectrum of muon neutrinos from the cern - sps accelerator . + transverse mass @xmath73 for the ( quasi-)two - body decays @xmath72 with @xmath107 ( left - hand column ) , @xmath108 ( middle column ) , and @xmath109 ( right - hand column ) . the unsmeared @xmath76 distributions for all events in each channel prior to any selections are shown in the top row . the smeared distributions for detected events are shown in the bottom row . [ exclude ] . + null - limit sensitivity to the @xmath4 transition ( at 90% c.l . ) of a 20-ton ebc detector deployed at 17 km from cern - sps and at 732 km from a muon storage ring with @xmath91 and 100 gev , assuming 10@xmath110 protons delivered by cern - sps and @xmath111 negative muons injected in a ring with a straight section of 25% . the shaded area on the right is the region of parameter space for @xmath4 suggested by a combined analysis of kamiokande and superkamiokande data @xcite . also illustrated are the best upper limits on @xmath85 for @xmath112 ev@xmath7 , as imposed by e531 @xcite and cdhs @xcite .

a conceptual scheme of a hybrid - emulsion spectrometer for investigating various channels of neutrino oscillations is proposed . the design emphasizes detection of @xmath0 leptons by detached vertices , reliable identification of electrons , and good spectrometry for all charged particles and photons . a distributed target is formed by layers of low - z material , emulsion - plastic - emulsion sheets , and air gaps in which @xmath0 decays are detected . the tracks of charged secondaries , including electrons , are momentum - analyzed by curvature in magnetic field using hits in successive thin layers of emulsion . the @xmath0 leptons are efficiently detected in all major decay channels , including @xmath1 . performance of a model spectrometer , that contains 3 tons of nuclear emulsion and 20 tons of passive material , is estimated for different experimental environments . when irradiated by the @xmath2 beam of a proton accelerator over a medium baseline of @xmath3 km / gev , the spectrometer will efficiently detect either the @xmath4 and @xmath5 transitions in the mass - difference region of @xmath6 ev@xmath7 , as suggested by the results of lsnd . when exposed to the neutrino beam of a muon storage ring over a long baseline of @xmath8 1020 km / gev , the model detector will efficiently probe the entire pattern of neutrino oscillations in the region @xmath910@xmath10 ev@xmath7 , as suggested by the data on atmospheric neutrinos . 1dec@xmath11 1dec@xmath12 _ 2pi@xmath13 1_3pi@xmath14 _ gg@xmath15 pacs numbers : 14.60.pq , 14.60.fg + keywords : neutrino oscillations , @xmath0 leptons , nuclear emulsion

the majority of the isotopes heavier than iron ( a@xmath356 ) are synthesized by neutron capture processes . the observed heavy elements distribution shows the presence of two main components , correlated to different nucleosynthetic processes : the s ( slow ) process and the r ( rapid ) process ( on the basis of the definitions given by @xcite in their pioneering work ) . the r process requires high neutron densities , and it is believed to occur during explosive phases of stellar evolution ( novae , supernovae and/or x - rays binaries ) . the s process is characterized by a slow neutron capture with respect to the corresponding @xmath4 decay : stable isotopes capture neutrons , while the radioactive ones decay ( @xmath5 or @xmath6 ) or capture a free electron . these isotopes are mainly created in the thermally pulsing agb ( tp - agb ) phase of low mass stars ( 1 @xmath7 m / m@xmath2 @xmath0 4 ) , where freshly synthesized elements are carried out to the surface by means of a recurrent mechanism called third dredge up ( tdu ) ( see @xcite and references therein ) . and solar metallicity .. ,width=321 ] in this phase the stellar structure consists of a partially degenerate carbon - oxygen core , an he shell separated from an h shell by the he - intershell region and by a convective envelope ( see fig.[mcore ] ) . the energy required to supply the surface irradiation is mainly provided by the h burning shell , located just below the inner border of the convective envelope . this situation is recurrently interrupted by the growing up of thermonuclear runaways , driven by violent he - burning ignitions . as a consequence of a thermal pulse ( tp ) , the region between the two shells ( he - intershell ) becomes unstable against convection ( for a short period ) , the external layers expand and , later on , the h shell burning temporarily dies down . in the he - intershell , he is partially converted into carbon . during the agb phase , main neutron sources are the @xmath8c(@xmath9,n)@xmath10o reaction , active in radiative layers during the interpulse period @xcite , and the @xmath11ne(@xmath9,n)@xmath12 mg reaction , marginally activated within the convective shell originated by the tp . in order to obtain a sufficient amount of @xmath8c for the activation of the s - process , a diffusion of protons from the h - rich envelope into the @xmath13c - rich radiative zone is needed : the diffused proton are captured from the abundant carbon via the @xmath13c(p,@xmath14)@xmath8n(@xmath5)@xmath8c nuclear chain , leading to the formation of a tiny @xmath8c - pocket . a major improvement in our stellar evolution code is the introduction of a physical algorithm for the treatment of the convective / radiative interface at the inner border of the convective envelope @xcite . during tdu episodes , the opacity of the envelope ( h - rich ) is significantly larger than the opacity of the underlying h - exhausted ( and he - rich ) region this fact causes an abrupt change of the temperature gradient at the inner border of the penetrating convective envelope . in this condition , the convective boundary becomes unstable , because any perturbation causing an excess of mixing immediately leads to an increase of the opacity and , in turn , to an increase of the temperature gradient . however , the steep pressure gradient that develops immediately below the formal border of the convective envelope limits the penetration of the instability , so that the average convective velocity should rapidly drop to zero . in order to mimic this behavior , we assume that in the region underlying the formal convective boundary , the average velocity follows an exponential decline , namely @xmath15 where _ d _ is the distance from the formal convective boundary , @xmath16 is the velocity of the most internal convective mesh , @xmath17 is the pressure scale height at the formal border of the convective envelope ( defined by the schwarzschild criterion ) and @xmath4 is a free parameter . the partial diffusion of protons in the top layers of the he shell gives naturally rise to the subsequent formation of a @xmath8c - rich tiny layer , the so - called @xmath8c pocket . in fig . [ tsol ] we plot the chemical profiles of trace isotopes in the region around the @xmath8c pocket . c pocket formed after the 3@xmath18 tdu episode ( @xmath19 model with solar metallicity).,width=321 ] the @xmath8c pocket ( solid line ) partially overlaps with a more external @xmath20n pocket ( long - dashed line ) . the maximum of the @xmath20n coincides with the region where the protons diffused from the convective envelope at the epoch of the tdu were abundant enough to allow a full cn cycle . this is also the region where the @xmath21na ( short - dashed - long - dashed line ) has been efficiently produced by the @xmath11ne(p,@xmath14)@xmath21na reaction . indeed , at the end of the previous tp , the mass fraction of @xmath11ne ( short - dashed line ) in the top layer of the he - rich intershell is of the order of 2@xmath22 and , around the @xmath20n maximum , this @xmath11ne is almost completely converted into @xmath21na , leading to the formation of a tiny @xmath21na pocket . + the introduction of the exponentially decaying profile of convective velocities automatically implies an assumption on the value of the free parameter @xmath4 : in the models presented in this work , we assume a value @xmath4=0.1 . in order to justify such a choice we perform some tests on a star with initial mass @xmath19 and different metallicities ( @xmath23=z@xmath24 and z=@xmath25 ) . we calculate the same sequences tp - interpulse - tp , starting from the first tp followed by tdu , and we evaluate the consequences of changing this parameter in the range [email protected] ( cristallo et al . , _ in preparation _ ) . while the amount of material interested by a single tdu episode increases by more than a factor of 3 from the case with no velocity profile ( @xmath4=0 ) and the extreme case ( @xmath4=0.2 ) , the @xmath8c@xmath27 , available for the heavy elements nucleosynthesis , grows up with increasing @xmath4 , reaches a maximum in correspondence of @xmath4=0.1 and then decreases down to negative values . we stress the fact that the net @xmath8c that contributes to the s - process nucleosynthesis is represented by @xmath8c@xmath27 , defined as the difference between the @xmath8c and the @xmath20n mass fractions ( @xmath20n is in fact a strong neutron poison via the @xmath20n(n , p)@xmath20c reaction ) . our result derives from a combination of two different physical processes : the growing of the @xmath8c@xmath27 is a direct consequence of the greater efficiency of the velocity profile algorithm , while its decreasing for large @xmath4 values is due to the fact that the convection efficiency in the zone interested by the velocity profile starts to be too high . this leads the most external layers of the he - intershell to be fully mixed and to consequently show an envelope - like h abundance , instead of the h profile needed for the formation of the @xmath8c pocket . we therefore propose an efficient way to calibrate the free parameter affecting the velocity profile algorithm , confining its possible values in a narrow range , and we adjust it in order to obtain the maximum @xmath8c pocket expected for the s - process nucleosynthesis . + the simultaneous solution of the stellar structure equations and a full network including all the relevant isotopes up to the termination point of the s - process path ( pb - bi ) requires a relevant computational power . for this reason , a post - process nucleosynthesis calculation , based on agb stellar models computed with a restricted nuclear network , was generally preferred @xcite . the coupling of a stellar code with a complete nuclear network has not been feasible so far , but this limitation has been overcome thanks to last generation of computers and to the adoption of smart algorithms to invert sparse matrixes of huge size . agb models presented in this work are therefore calculated by using an extended set of nuclear processes including all chemical species involved in the s - process nucleosynthesis @xcite . the models presented here have been obtained by including into the franec stellar evolution code @xcite a full nuclear network , containing about 500 isotopes ( from h to the pb - bi - po ending point ) linked by more than 750 reactions . reaction rates of isotopes involving charged particles are generally taken from the nacre compilation @xcite , while the neutron captures cross sections are mainly taken from @xcite . weak - interaction rates ( electron captures , and @xmath4 decays ) are interpolated as a function of the temperature and electron density , while at temperatures lower than @xmath28 k we assume a constant value equal to the terrestrial one . the presence of isomeric states of unthermalized isotopes , which lead to ramifications of the s - process flux , is followed in detail . this network is continuously upgraded according to the latest theoretical and experimental nuclear physics improvements . the franec models benefit of another important improvement , whose effects are particularly important at low metallicities : the introduction of c - enhanced low temperature opacities . in principle , a stellar evolutionary code should account for the variations of the envelope chemical composition ; in practice , only variations of the main constituents ( h and he ) are usually considered . nevertheless , in the atmosphere of an agb star , both c dredge - up and the possible conversion of c into n ( by cn cycle ) substantially affect the molecular contribution to the opacity for temperatures lower than 4000 k. this is due to the fact that , when a carbon excess is present ( i.e. when the number ratio c / o@xmath291 ) , carbon bearing molecules ( c@xmath30 , cn , c@xmath30h@xmath30 , c@xmath31 ) start forming and their contribute to opacity become dominant . therefore , in order to overcome this limitation , we calculate low temperature c- and n - enhanced opacity tables , suitable for agb model calculations @xcite . the new agb models succeed in reproducing the photometric and spectroscopic properties of their observational counterparts ( see section [ lowz ] ) . and solar metallicity after the fifth tdu ( dotted curve ) and the last tdu ( solid curve).,width=321 ] the final c / o envelope ratio in the solar metallicity model is 1.88 , while it becomes larger than 1 when the core mass is about 0.6 @xmath32 and the bolometric magnitude is about @xmath335.0 mag , in good agreement with the galactic c - star luminosity function @xcite . a @xmath8c pocket , whose mass extension decreases pulse after pulse , forms after each tdu episode . the @xmath8c in the first pocket is only partially burnt during the interpulse and the residual is engulfed into the convective zone generated by the subsequent tp . in this case , the @xmath8c burning takes place at the bottom of the convective shell ( @xmath34 k ) producing a very high neutron density , this fact implying interesting peculiarities on the following nucleosynthesis ( see @xcite ) . in particular , some branchings , which remain closed during a standard radiative @xmath8c burning and are marginally activated during the @xmath11ne(@xmath9,n)@xmath12 mg burning , are now open : in a very short temporal step ( @xmath35 years ) we obtain a consistent production of neutron - rich isotopes normally by - passed by the standard radiative @xmath8c s - process , among which @xmath36fe , @xmath37kr , @xmath38rb and @xmath39zr . concerning the standard radiative s - process nucleosynthesis , all the elements and z=@xmath40 ( left panel ) and z=@xmath41 ( right panel).,width=453 ] ( from sr to pb ) result enhanced . we find that the abundance of sr , y and zr at the first s peak , the so - called ls elements ( _ light _ s elements ) , is comparable with the one of the hs elements ( _ heavy _ s elements ) ba , la , ce , pr , nd at the second s peak ( see fig . [ disol ] ) . lead is under - produced with respect barium . the [ hs / ls]=log(@xmath42(a)/@xmath42(b))@xmath33log(@xmath42(a)/@xmath42(b))@xmath24 . ] attained in the envelope when c / o=1 is in good agreement with those measured in galactic c(n ) giants @xcite . the @xmath8c(@xmath9,n)@xmath10o reaction ( the main source of neutrons ) is primary - like , i.e. not directly affected by the metallicity of the pristine material . however , since the build - up of heavy nuclei requires neutron captures starting from fe seeds , the s - process is expected to decline with metallicity ( it is therefore of secondary nature ) . nevertheless , while the iron seeds scale with the metallicity , when decreasing the number of iron seeds the number of neutrons available per seed is larger . the dependence of s - process yields at different metallicities is therefore very complex and non linear ( see e.g. @xcite ) . at metallicities z@xmath43 the s - process distribution is peaked around the hs elements ( see fig . [ di_int ] ) . moreover , we obtain an increase in the relative production of carbon and fluorine with decreasing the metallicity : this is a direct consequences of the greater efficiency of the tdu episodes , when the convective envelope penetrates into the @xmath13c - rich he - intershell ( see @xcite ) . we want to stress the fact that the production of @xmath13c is of primary nature ( it is in fact produced by the activation of the triple @xmath9 reaction in the he - intershell ) . the nature of the fluorine production , that basically depends on the nucleosynthesis of @xmath44n ( see e.g. and z=@xmath45 ( left panel ) and z=@xmath25 ( right panel).,width=453 ] @xcite ) is more complex , being partly of primary nature ( due to the contribution of the @xmath8c pocket ) and partly of secondary nature ( due to the contribution of the ashes of the h - burning shell ) . finally , let us note that at these metallicities the production of lead ( mainly @xmath46pb ) starts being very efficient . at low metallicities , most of the seeds are converted into @xmath46pb , at the termination point of the s - process fluency ( see e.g. @xcite ) . trends illustrated in fig . [ di_low ] ( which relates to the @xmath4710@xmath48 and to the @xmath4710@xmath49 models ) confirm the previous sentence : a consistent lead production ( @xmath23=82 ) is in fact found in both models . as already stressed in the previous section , the lower the metallicity is , the larger surface carbon overabundances are found : [ c / fe]@xmath502.2 at @xmath23=1@xmath5110@xmath48 and [ c / fe]@xmath503.0 at @xmath23=1@xmath5110@xmath49 . + low metallicity stars enriched in s - process elements show consistent enhancements in the hs region , spanning in the range 0.8@xmath0[hs / fe]@xmath02.3 . even if our final [ hs / fe ] value lays within the observed range , we can not reproduce the observed spread with a single evolutionary model . in spite of this , we want to point out the importance of the opacity coefficients on the stellar evolution of low metallicity models . as already outlined in section [ opa ] , in our models the opacity change caused by the variation of the internal chemical composition has been taken into account by linearly interpolating between tables with different carbon , nitrogen and helium abudances @xcite . however , the procedure followed by the majority of modelers is to adopt the opacity coefficients calculated at the initial metallicity and to use them for the entire evolution . in order to evaluate the possible effects induced by these different opacity treatments , we compute another model at @xmath23=1@xmath5110@xmath49 by using for the entire evolution the opacity coefficients calculated for the initial metallicity ( we refer to this calculation as the _ z - fixed _ case ) and we compare it with respect to the @xmath23=1@xmath5110@xmath49 model already presented ( _ standard _ case ) . the tp - agb phase of the _ z - fixed _ case is longer ( a factor of 2 more than the _ standard _ model ) , this implying a natural increase in the number of tps followed by tdu ( n=49 with respect to the 15 found in the _ standard _ model ) . consequently , the surface abundances of the _ z - fixed _ case are larger with respect to the _ standard _ case ( see fig . [ compall ] ) , showing a better agreement with observations . we argue that these difference in the surface distributions are due to the mass loss rate , which results and z=@xmath25 . the solid line refers to our _ standard _ model ; the dashed line refers to a model with the same initial inputs but with a different opacity treatment ( _ z - fixed _ model ) . see text for details.,width=321 ] less efficient in the _ z - fixed _ model with respect to the _ standard _ case , due to the fact that the surface temperature is always larger and the radius is less expanded @xcite . as a consequence of so different mass - loss rate histories , in the _ z - fixed _ case we find a large increase of the total dredged up material during the agb phase with respect to the _ standard _ case and , therefore , the resulting final heavy element enhancements result larger . our results seem therefore to suggest that the model computed with opacity tables at fixed metallicity is more efficient in reproducing spectroscopic data . there are however observational counterparts supporting the opposite conclusion , driven by the study of c - stars in dwarf spheroidal galaxies ( dsgs ) ; we refer in particular to the draco dwarf spheroidal galaxy @xcite . the observed giants belonging to this dsph present surface temperatures in the range 3.55@[email protected] . note that this range is never attained by the _ z - fixed _ case , while it is entirely spanned by the _ standard _ case ( see fig.4 in @xcite ) . + another important aspect which has to be addressed is the choice of the mass - loss history . actually , observation can not lead modelers in parameterizing the mass - loss at low metallicities , basically due to the fact that up to date a clear determination of the gas - to - dust ratio ( needed to properly calculate the mass - loss ) has not been feasible at non - solar metallicities @xcite . for this reason we adopt the same mass - loss history at all metallicities ( see @xcite for a description of our choice ) . preliminary results show however that when using a different ( milder ) mass - loss @xcite the surface overabundances increase on average for 0.2 dex ( cristallo et al . _ in preparation _ ) . + another source of uncertainty in modelling agb stars at low metallicity is the choice of the mixing length parameter , which has been calibrated by calculating a standard solar model . if using initial abundances from @xcite , we fit the observed properties of the sun by using a value of the mixing length parameter @xmath52=2.1 ( see @xcite ) . due to the current lack of indications for the assumption of a different @xmath52 value at low metallicities , we adopt the same value at all metallicities : preliminary results indicate however that when using a lower value ( i.e. @xmath52=1.8 ) the surface overabundances decrease on average for 0.2 dex ( cristallo et al , _ in preparation _ ) . in the case of very metal poor agbs , we find that , for a given metallicity , there exists a lower mass limit for which normal agb evolution may occur @xcite . stars with smaller initial mass experience protons engulfment episodes during the first fully developed tp ( see e.g. @xcite ) , which lead to a peculiar s - process nucleosynthesis , a low @xmath13c/@xmath8c and a significant synthesis of primary n. owing to the low entropy barrier , protons are engulfed within the he - intershell , where the high temperature and the large c abundance induce a violent h - burning flash . the consequent synthesis of @xmath8c allows the activation of the @xmath8c(@xmath9,n)@xmath10o reactions and a very large neutron density is attained . thus , the production of neutron - rich nuclei , normally by - passed by a standard s - process , can occur in these stars . later on , when the intershell cools down , a deep tdu occurs . since in the envelope the c / o becomes grater than 1 , the large n abundance favors the formation of cn molecules , thus inducing a further increase of the radiative opacities . this new set of very low metallicity models , that represents the current frontier of agb modelling , is still under analysis . the present work demonstrates that , nowadays , the computational power allows the coupling between a stellar evolutionary code and a full nuclear network . a different treatment of the internal border of the convective envelope with respect to a bare schwarzschild criterion allows the formation of the so - called @xmath8c pocket . for the first time in the literature , it has been possible to directly compare theoretical models and observational data both from a physical ( luminosities , surface temperatures ) and a chemical ( light elements and heavy elements abundances ) point of view . moreover , we furnish a uniform set of yields ( from hydrogen to lead ) at different metallicities . the importance of the adopted mass loss rate and of molecules contribution to opacity has been pointed out . while the first problem is still under analysis , the second one has been solved by using opacity tables with enhanced carbon and nitrogen abundances . finally , we have addressed the importance of very low metallicity agb models , by describing first preliminary ( and promising ) results . i sincerely thanks prof . o. straniero and prof . r. gallino for their invaluable expertise and their deep knowledge of the arguments treated in this paper . without a continuous interaction and ideas sharing with them i would not acquire the needed scientific uprightness to make research in a proper way .

people usually smile when astrophysicists assert that we are _ sons of the stars _ , but human life confirms this sentence : about 65% of the mass of our body is made up of oxygen , carbon occurs in all organic life and is the basis of organic chemistry , nitrogen is an essential part of amino acids and nucleic acids , calcium is a major component of our bones . moreover , phosphorus plays a major role in biological molecules such as dna and rna ( where the chemical codes of life is written ) and our blood carries oxygen to tissues by means of the hemoglobin ( an iron pigment of red blood cells ) . all these elements have been created in stars . i just list some examples related to human body , but also common element such as aluminum , nickel , gold , silver and lead come from a pristine generation of stars . the abundances in the solar system are in fact due to the mixing of material ejected from stars that polluted the universe in different epochs before the sun formation , occurred about 5 billion years ago , after the gravitational contraction of the proto - solar cloud . low mass agb stars ( 1@xmath0@xmath1/m@xmath2@xmath03 ) are among the most important polluters of the milky way , because of the strong winds eroding their chemically enriched envelopes . they are responsible for the nucleosynthesis of the main component of the cosmic s - elements . [ 1999/12/01 v1.4c il nuovo cimento ]

the future of optical observational astronomy looks bright , with excellent prospects for the advent of larger aperture telescopes both on the ground and in space . as the telescope size increases , so does the cost and complexity of a project . the 2010 decadal survey will therefore face difficult choices for ground- and space - based astronomy investments in the coming decade . to facilitate these choices , it is important to understand the relative advantages of ground- and space - based facilities for observational astronomy . in particular , the ever increasing capabilities of adaptive optics ( ao ) from the ground motivate a critical side - by - side comparison of the observational parameter space accessible to either type of facility.=-2 we focus here mostly on comparison of the capabilities of a @xmath0 m ground - based telescope with an 816 m space - based telescope . examples of the former are the thirty meter telescope ( tmt ; stone et al . 2009 ) and the giant magellan telescope ( gmt ; mccarthy et al . examples of the latter are the advanced - technology large - aperture space telescope ( atlast ; postman et al . 2009 ) and the terrestrial planet finder coronagraph ( tpf - c ; levine et al . we compare only briefly the capabilities of current ground - based 810 m class telescopes to smaller space - based telescopes such as the 2.4 m hubble space telescope ( hst ) , the 6.5 m james webb space telescope ( jdem ) , or the proposed joint dark energy mission ( jdem / idecs ; gehrels et al . 2009 ) . we restrict attention to the regime below @xmath2 m , for two reasons . first , this is where future large space - based proposals are likely to focus their attention ; jwst will already provide an important new large aperture facility for science longward of @xmath2 m . second , below @xmath2 m is where the future capabilities of ao from the ground are most uncertain and least documented . ao corrections in the infrared ( ir ) are typically more capable and less technically challenging than in the optical ( see section [ s : aotypes ] ) . however , in comparing performance of space versus ground in the ir , careful consideration must be given to the role of background emission in achieving a given signal - to - noise ratio @xmath3 ( see section [ s : sn ] ) . we will not discuss here the relative cost for new optical observing facilities on the ground or in space . observational facilities in space are generally more expensive to build and operate than those on the ground . so they are usually pursued only if they open up parts of parameter space that are not accessible from the ground . we highlight which parts of observational parameter space are uniquely accessible only from space . scientific motivations for access to this parameter space are briefly mentioned where relevant . costs and science drivers are discussed in detail in other submissions to the decadal survey . one of the main advantages for ground - based facilities is that they can generally be constructed with larger telescope diameters @xmath4 than what is possible in space , which impacts two important characteristics : the collecting area and the diffraction limit . [ [ collecting - area ] ] 2.1.1 . collecting area + + + + + + + + + + + + + + + + + + + + + + the collecting area of a telescope scales as @xmath5 . this implies that large - aperture ground - based telescopes are generally able to collect more photons than space - based telescopes , which is one of the factors that determines the achievable depth and signal - to - noise ratio ( @xmath3 ; see section [ s : sn ] ) . [ [ diffraction - limit ] ] 2.1.2 . diffraction limit + + + + + + + + + + + + + + + + + + + + + + + + the diffraction limit of a telescope scales as @xmath6 . this implies that large - aperture ground - based telescopes generally have a smaller diffraction limit than space - based telescopes . however , due to atmospheric turbulence it is more difficult for ground - based telescopes to achieve image quality near the diffraction limit than it is for space - based telescopes , especially at optical wavelengths ( see sections [ ss : imqual ] , [ s : aotypes ] and [ s : contrast ] ) . one of the most fundamental advantages of space - based over ground - based observations is the fact that light can be observed unaffected by the atmosphere . this impacts several important characteristics of the observational parameter space . [ [ accessible - wavelength - range - and - absorption ] ] 2.2.1 . accessible wavelength range and absorption + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + ultra - violet observations are not possible at all from the ground , due to atmospheric absorption . this blocks an entire wavelength range in which many astrophysical problems can be uniquely studied . even in the optical there are spectroscopic absorption features ( possibly time - variable ) associated with telluric bands . these need to be corrected using observations of standard stars , which themselves are only characterized to finite accuracy . [ [ background ] ] 2.2.2 . background + + + + + + + + + + + + + + + + + the atmosphere creates a background in ground - based observations that needs to be subtracted . figure 1 compares the background at mauna kea and l2 . the shot noise from the background affects the the achievable depth and @xmath3 ( see section [ s : sn ] ) . in addition , background variations can lead to systematic errors due to imperfect subtraction.=-2 [ [ line - emission ] ] 2.2.3 . line emission + + + + + + + + + + + + + + + + + + + + in addition to a continuum background , ground - based observations accumulate photons from atmospheric emission in well - defined bands with many narrow emission lines ( see figure 1 ) . the presence of these emission lines increases towards the ir , and their strengths vary with time . this limits the ability to study spectral features in astronomical objects at particular wavelengths . [ [ atmospheric - extinction ] ] 2.2.4 . atmospheric extinction + + + + + + + + + + + + + + + + + + + + + + + + + + + + + atmospheric extinction must be corrected on the basis of airmass estimates and standard star observations , to obtain absolute photometry . the accuracy with which these corrections can be done are critical for certain areas of science . the image quality depends both on the diffraction limit of the telescope ( section 2.1.2 ) and the ability of the telescope to reach this diffraction limit . [ [ strehl - ratio - sr ] ] 2.3.1 . strehl ratio ( sr ) + + + + + + + + + + + + + + + + + + + + + + + + the sr is the ratio of the peak flux in the normalized point - spread function ( psf ) to that for a diffraction limited system . it ( and other measures of image quality ) describe the extent to which the light is concentrated in the psf core as opposed to the wings . a space - based telescope is typically designed to be diffraction limited ( commonly defined as @xmath7% ) , at some target wavelength driven by the science . by contrast , large ground - based telescopes need an ao system that aims to optimize a combination of the complimentary goals of high sr and large field of view ( see sections [ s : aotypes ] and [ s : contrast ] ) . even when an ao system achieves a diffraction - limed core , a low sr can severely limit the science . for example , in a crowded field the psf wings of bright stars create an elevated background that drowns out the light of fainter stars . [ [ field - of - view - fov ] ] 2.3.2 . field of view ( fov ) + + + + + + + + + + + + + + + + + + + + + + + + + + the fov sizes for space- and ground - based telescopes are both limited by technological , design , and cost constraints related to their optics and detectors . however , for ground - based observations the fov is further limited by the area over which ao correction is possible ( see section [ s : aotypes ] ) . this limits the ability to do certain kinds of science , e.g , wide area surveys at high - resolution . [ [ stability ] ] 2.3.3 . stability + + + + + + + + + + + + + + + + another advantage of space observations is the long - term stability of the environment , with no gravity , no seismic disturbances , no weather , and generally smaller thermal fluctuations . as a result , the psf and optical geometric distortions are extremely stable . this makes the space - environment uniquely suited for science that requires high photometric or geometric precision . absolute photometry is possible from space to levels around @xmath8 mag , and relative differential photometry to levels of order @xmath9 mag . the latter is critical for photometric variation studies such as for planet transits ( as in the kepler mission ) . the photometric quality for ground - based ao is limited by temporal and spatial ( field - dependent ) psf variations , which become more significant for decreasing wavelength and increasing field size . in the near - ir , current accuracies of @xmath10 mag ( e.g. , davidge et al . 2003 ; vacca et al . 2007 ) can likely be reduced to near space quality with the advent of , e.g. , mcao systems ( see section [ s : aotypes ] ) . ground - based ao observations ( e.g. , for the galactic center ) and space - based observations ( e.g. , for globular cluster stars ) have both produced superb proper motions results , with the former benefitting from the small diffraction - limited core , and the latter from the exquisite geometric stability over large time baselines.=-2 [ [ time - sampling ] ] 2.4.1 . time sampling + + + + + + + + + + + + + + + + + + + + one other obvious advantage of space , and in particular the l2 sun - earth lagrange point , is that there is no day - night cycle ( this does not apply to low - earth orbit satellites such as hst ) . from l2 it is possible to continuously follow time - variable phenomena and to take continuous deep exposures , without the need to compare or co - add data from different nights during which observing conditions may have been different . [ [ sky - coverage ] ] 2.4.2 . sky coverage + + + + + + + + + + + + + + + + + + + ao observations require bright guide stars to correct for atmospheric turbulence . with natural guide stars ( ngss ) , only @xmath11% of the sky is accessible depending on the galactic latitude , for k - band imaging at moderate strehl ratios ( @xmath12% ; frogel et al . this limitation has now been mostly overcome through the use of laser guide stars ( lgss ) . however , even with an lgs system the sky coverage is not complete , since it still requires an ngs for the tip - tilt correction ( although this ngs can be several magnitudes fainter than for a pure ngs system ) . in general , sky coverage drops with increasing sr and galactic latitude . for example , nfiraos on tmt is expected to cover @xmath13 at the galactic pole ( ellerbroek , priv . comm . ) and up to 100% at low galactic latitude . projects and proposals exist to improve sky coverage further using various techniques , both on tmt and on smaller telescopes ( e.g. , keck , cfht ) . the limitations imposed by atmospheric turbulence can be characterized by the fried diameter @xmath14 , turbulence lifetime @xmath15 , and isoplanetic angle @xmath16 . all are proportional to @xmath17 . for example , typical magnitudes for a good groundbased site are @xmath18 cm , @xmath19 ms , and @xmath20 at 5000 and @xmath21 cm , @xmath22 ms , and @xmath23 at @xmath24 m . for a given sr , the number of subapertures that must be corrected is proportional to @xmath25 in a time interval @xmath15 . consequently the technology challenges of both applying a sufficiently high fidelity wavefront correction , and making a sufficiently accurate wavefront measurement within a turbulent cell of scale length @xmath14 in a time interval @xmath15 are much more manageable for ao corrections in the near - ir than at optical wavelengths . the number of photons required for a constant level of correction is roughly @xmath26 ( the @xmath27 comes from the area of the subaperture and time scale over which a given number of photons are required ; the @xmath28 comes from the need to achieve the same path error expressed in waves ) . hence , lower wavelengths require much larger laser power for wavefront sensing.=-2 today , ao for near - ir observations is in routine operation on the keck , vlt , gemini and other telescopes . a detailed overview of current ao capabilities was provided by frogel et al . ( 2008 , 2009 ) , together with a roadmap for future development . they found that both ao performance and the number of ao - enabled refereed science papers has grown steadily over the last five years . we discuss the various types of ao in turn below . [ [ classical - ao ] ] classical ao + + + + + + + + + + + + in classical ao , a single guide star , either natural ( ngs ) or laser ( lgs ) , is used to measure the deformations of the incoming wavefront with a wavefront - sensor and to correct them with the help of a single deformable mirror ( dm ) . in this case the fov is limited to a few times the isoplanetic angle @xmath16 . since @xmath29 for @xmath30 m , this is an important limit for many scientific applications . one limitation of classical ao with an lgs is the so - called called `` cone effect '' . this results from the incomplete sampling of the turbulence in front of the aperture due to the finite distance of the lgs . this becomes increasingly problematic for larger telescope diameters and shorter wavelengths . [ [ ltao ] ] ltao + + + + in laser tomography adaptive optics ( ltao ) , multiple lgss and wavefront sensors are used to measure the full volume of turbulence above the telescope in order to solve for the cone effect and operate a single conjugate deformable mirror , as in classical ao . this is what is planned for visible 10m - class ao systems , and for near - ir first light use on 30m - class ground - based telescopes such as the tmt . the sr will range from 0.3 at @xmath31 m to 0.8 at @xmath32 m over a fov of about @xmath33 , with a technical fov for guide star acquisition of @xmath34 . the short wave cutoff for science will be @xmath35 m , where one might still expect a diffraction limited core at very low sr ( ellerbroek , priv . comm . ) . [ [ mcao ] ] mcao + + + + an exciting recent advance has been the advent of multi - conjugate adaptive optics ( mcao ) , in which multiple dms are optically conjugated at different altitudes . multiple wavefront sensors use lgss to tomographically measure and compensate for turbulence - induced phase aberrations in three dimensions . this new technique increases the compensated fov , provides a more uniform psf over the field , and also solves for the `` cone effect '' . the solar community has been using mcao to deliver arc - minute scale , fully corrected near - ir images of the solar granulation . a demonstrator at the eso / vlt has delivered stable arc - minute scale images with @xmath36% psf variations . the first mcao on gemini will deliver nearly - uniform performance over a @xmath34 fov with sr ranging from 45% to 80% from 1@xmath37 m . [ [ glao ] ] glao + + + + ground - layer adaptive optics ( glao ) is a `` lite '' form of mcao , in which only the ground - layer turbulence is corrected . this technique does not provide the same image quality improvement as mcao . it is therefore sometimes referred to as `` seeing enhancement '' , with fwhm improvements of a factor of a few . however , the technique offers the advantage of a large fov ( up to several arcmin ) . moreover , it has the potential to be implemented on 30m - class ground - based telescopes at visible wavelengths . [ [ moao ] ] moao + + + + in multi - object adaptive optics ( moao ) , a number of objects are selected in the field . each object goes through it own ao system with one dm per object . the correction is based on tomographic knowledge and is open - loop . like mcao , moao provides the potential for significant fov enhancement over the isoplanetic angle . its concepts are well developed and have been run on a testbed , but remain to be demonstrated on large astronomical telescopes.=-2 [ [ visible - light - ao ] ] visible - light ao + + + + + + + + + + + + + + + + even in the optical , improvements in technology have demonstrated correction around bright sources across a narrow ( 1@xmath38 ) fov . these advances have so far been led by investments by the us air force on the 3.5 m sor telescope . in the next decade we should see these technologies ( and others pursued on , e.g. , cfht / vasao and palomar / palm3k ) move to large ground - based telescopes , as the cost - effectiveness of high - density dms and high - power lasers improve ( e.g. , dekany et al . however , the delivered srs will always decrease strongly towards lower wavelengths . moreover , the technological challenges to extend the fov remain enormous . because the guide star brightness requirement scales with @xmath39 , challenging amounts of laser power will be necessary for visible ao ( independent of @xmath4 ) . solutions exist to mitigate these power requirements , for example using predictive control or uplink correction ( gavel et al . laser power drives the cost of these systems . routine visible - light ao operation with _ both significant sr and fov _ on 30m - class ground - based telescopes is not foreseen in the coming decade(s).=-2 [ [ exao ] ] exao + + + + in extreme ao ( exao ) , narrow field , high - sr systems are specifically designed to detect planets around nearby bright stars , as discussed in section [ s : contrast ] . [ [ wavefront - control - in - space ] ] wavefront control in space + + + + + + + + + + + + + + + + + + + + + + + + + + although space has no turbulent atmosphere to correct for , figure control is definitely important . following in the footsteps of jwst , future observatories are increasingly likely to use wavefront - sensing for optimum image quality . although corrections can be done at much lower speeds that from the ground , the necessary technologies ( e.g. , mems ) have strong overlap with those ( being ) developed from ground - based ao.=-2 a crucial consideration in the comparison between ground - based and space - based observations is the atmospheric sky background emission @xmath40 ( see figure 1 ) , which becomes particularly important in observations designed to reach the faintest possible limits . the noise contribution in imaging observations is determined by the photometric aperture size @xmath41 . for seeing - limited observations @xmath42 , independent of @xmath4 , while for observations with a diffraction limited core @xmath43 . for given target flux and wavelength or passband , this yields for the limit in which background noise dominates other sources of noise that @xmath44 for seeing - limited or diffraction - limited observations , respectively . here @xmath45 is the product of the throughput and quantum efficiency of the system , which nowadays is close to unity for both ground- and space - based systems . these equations make it clear that the deepest science benefits are derived from larger telescopes , better image quality , and lower background . = 0.47=0.525 the background flux ( from atmospheric molecular , ionic and continuum emission and telescope thermal emission ) for a ground - based telescope on mauna kea as compared to a space - based telescope at l2 . + + * table 1 ( right ) : * sensitivity comparison between space - based ( l2 ) and ground - based observatories . the imaging _ observing time in hours _ is listed to reach @xmath46 at the magnitude listed in the left column . see table notes and main text for details on the calculations . for the ground - based 30 m , each line uses a different psf , as listed in the last column . for the @xmath47-band , this spans the range of what may be achievable . for extremely faint sources ( e.g. , @xmath48 ) , the atmospheric sky background emission , even at the best ground - based sites , is at least @xmath49@xmath50 times brighter than the source . to illustrate what is feasible in this context , table 1 shows the exposure times needed to reach a given @xmath3 for a set of representative target magnitudes . the exposure times are given for the @xmath47 and @xmath51-bands at @xmath52 and @xmath53 m , respectively . this brackets the range of wavelengths on which we focus in the present paper . these values were obtained from the integration or exposure time calculators for 2.4 m ( hst ) , 8 m and 16 m space - based observatories , the latter two located at l2 , as well as 8 m and 30 m ground - based observatories with ao capability . for the ground - based 8 m we used existing capabilities , in particular gemini / niri+altair in the @xmath51-band and a gaussian fwhm of @xmath54 ( good seeing ) in the @xmath47-band ( at present , there are no general - user optical ao systems available on 8 m class telescopes ) . for the ground - based 30 m telescope in the @xmath51-band we optimistically assumed diffraction - limited performance with the same sr as for the space - based systems . for the @xmath47-band we present a sampling of three potential psfs that span the range of where we might be in a decade or two : ( a ) a gaussian fwhm of @xmath54 ( good seeing ) ; ( b ) a gaussian fwhm of @xmath55 ( optimistic estimate with a successful advent of visible glao ) ; or ( c ) a diffraction limited core with sr = 10% ( optimistic estimate with a successful advent of visible ltao , mcao , or moao).=-2 the results for the current generation of facilities show that in the j - band , hst ( with the wfc3 camera ) goes several magnitudes deeper than a ground - based 8 m , with only a moderate difference in spatial resolution . in the v - band , both facilities can reach comparable depths , due to the smaller difference in background at @xmath56 ( see figure 1 ) . however , hst provides the better spatial resolution . for future facilities , a 16 m space - based observatory at l2 will be capable of reaching @xmath57 and @xmath58 in integrations shorter than 1 day . this is a completely unexplored parameter space . a 30 m ground - based telescope will fall short of these limits by 24 mag . the difference is due to a combination of a higher sky emission and ( in @xmath47 ) poorer image - quality on the ground . an 8 m telescope in space at l2 still goes deeper in @xmath47 than a 30 m ground - based telescope by 13 mag . however , in @xmath51 the depth is more similar and the ground - based telescope with diffraction - limited ao will have the better spatial resolution . figure 2 provides a graphical way to look at results of calculations of this nature as function of wavelength , with the spectrum of the earth as it would be seen at 20pc overplotted.=-2 = 0.45=0.45 10-sigma point source sensitivities ( njy ) for background - limited 1-hour broadband ( r=5 ) imaging for : an 8-m and 16-m space telescope ( black ) ; a 30-m ground - based telescope with diffraction - limited ao ( red ) ; and jwst ( dark blue ) . sensitivities in this figure were calculated with the methodologies of beckwith ( 2008 ) . the superposed spectrum ( light blue ) shows for comparison a terrestrial exoplanet at a distance of 20 pc . to detect such a planet one also has to contend with additional backgrounds not shown here : ( 1 ) the exo - zodi light around the parent star ( a factor @xmath59 larger than the l2 background itself ) ; and ( 2 ) the psf halo of the parent star ( depends on achieved contrast ; see section [ s : contrast ] ) . the band at the bottom of the plot shows the ground - based sky absorption ( black = sky opaque to external radiation ) . + * figure 3 ( right ) : * ratio of exposure times for reaching @xmath46 as function of wavelength @xmath28 , with a ground - based 30 m telescope or an l2 space - based telescope of size 16 m ( top ) or 8 m ( bottom ) , respectively . ratios are indicated for four different spectral resolutions @xmath60 , with the top curve in each panel being the lowest resolution . similar considerations apply to spectroscopic observations , since the spatial resolution determines the smallest slit width and spatial aperture dimension for producing a final summed spectrum . as an example , figure 3 compares the time to reach @xmath61 for a point source as a function of spectral resolution for a 30-m ground - based telescope , versus the time required for a space - based 8 m and 16 m observatory located at l2 . as in table 1 , all are assumed to have the same instrument and detector performance . in these calculations we have assumed that the 30 m ground - based telescope is diffraction - limited at wavelengths longer than @xmath2 m , and is seeing limited ( @xmath54 fwhm ) at lower wavelengths . some form of optical ao would decrease the advantage of the space - based observatory , as in table 1.=-2 figure 3 shows that an 8 m space - based telescope is 10 to 100 times faster than a ground - based 30 m for all optical seeing - limited imaging ( @xmath62 ) and up to 40 times faster for most low - resolution ( @xmath63 ) spectroscopy . the space - based 8 m is also more sensitive for medium - resolution optical spectroscopy ( @xmath64 ) . similarly , a 16 m space telescope is much faster for all spectroscopy in the visible compared to a seeing - limited 30 m ground - based telescope . wide - field optical imaging and moderate resolution spectroscopy at the depth allowed by a new large space - based telescope will allow important new studies of , e.g. , planets ( kasting et al . 2009 ) , distant and/or faint galaxies ( giavalisco et al . 2009 ) , and resolved stars outside the local group ( brown et al . 2009 ) . at highest spectral resolutions , background noise ceases to be the dominant noise contribution , so that space observatories loose their edge . also , another important metric is how many sources can be observed spectroscopically at any given time . such multiplexing is possible in space ( e.g. , jwst / nirspec ) , but ground - based observatories may well have fewer constraints in pushing this to its limits . [ [ exao-1 ] ] exao + + + + extreme adaptive optics pushes high - contrast observations , geared in particular towards study of exoplanets . it gives a very high correction ( @xmath65% ) within a small fov ( a few arcsec , corresponding to a few aus to a few tens of aus around nearby stars ) . the target star itself is used for wavefront sensing , therefore eliminating any anisoplanetic errors . exao requires large numbers of corrected modes , and fast correction rates in order to achieve the very high image quality . currently , exao systems are designed and built for the near infrared where the correction is easier than in the visible ( see section [ s : aotypes ] ) . [ [ coronagraphs - and - high - contrast - calibration ] ] coronagraphs and high - contrast calibration + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in order to produce high - contrast images , exao needs to be combined with coronagraphs and additional calibration schemes ( active calibration , speckle nulling , differential imaging , etc . ) . many coronagraphs have been developed and proven in the laboratory . the best laboratory experiments have reached @xmath66 contrast in medium bands ( 10% ) , which proves the feasibility of the concept for the detection of earth analogs . coronagraphs work best for perfect non - obstructed apertures . although some schemes exist to mitigate the effects of a central obstruction or segmentation , the performance is severely affected by the diffraction of these geometrical features , especially with the typically large central obstruction and wide support structures required on future 30m - class ground - based telescopes ( hereafter referred to as extremely large telescopes , elts ) . it is not clear whether internal coronagraphs can be designed to deliver @xmath1 contrast with large on - axis segmented telescopes , but the task is facilitated by the increased angular resolution of elts . the coronagraphs need only reach this contrast level at 1020 resolution elements ( @xmath6 ) for the detection of an earth - twin , as opposed to a few for smaller telescope . after coronagraphy , space - observations and ground+exao - observations require similar levels of calibration , in order to improve the performance and remove residual starlight propagation artifacts ( speckles ) . this task is made easier in space because of the greater thermal and mechanical stability of a space - based telescope . [ [ state - of - the - art ] ] state of the art + + + + + + + + + + + + + + + + on the ground , new instruments are being built for current 8m - class telescopes and will start operating in 2011 . these instruments ( e.g. , gpi , sphere , subaru , palomar ) involve exao , advanced coronagraphy , spectrographs and polarimeters . these projects will focus on observations of young ( @xmath67 gyr ) giant planets in the near - ir ( 0.9@xmath37 m ) . elts also envision high - contrast instruments , but very little funding has been invested so far . their ao is more challenging , and mirror segmentation increases the difficulty for coronagraphic efficiency and stability . the science goals for high - contrast ground - based elts are very exciting and mostly include the study of young planets in star forming regions , mature reflected light giant planets , known radial velocity ( rv ) planets , neptunes for nearby stars , and high - resolution images of protoplanetary disks . [ [ science - drivers ] ] science drivers + + + + + + + + + + + + + + + imaging and spectroscopic characterization of exoplanets or disks is the principal science motivation for high - contrast observations . the most exciting goal is arguably the direct detection of habitable terrestrial planets and the search for spectroscopic biomarkers . strong motivations exist for doing these observations at short wavelengths ( @xmath68 m ) , as discussed in , e.g. , kasting et al . ( 2009 ) and lawson et al . this is because of the strong oxygen band at 760 nm , as well as other potential biomarkers ( ozone , vegetation red edge , rayleigh scattering ) . this prompts the question whether such a science program could be achieved with a ground - based 30 m class telescope . we use the sun - earth system at 10 pc as a template for addressing this discussion . [ [ idealized - analysis ] ] idealized analysis + + + + + + + + + + + + + + + + + + adaptive optics is fundamentally limited by the capacity to analyze the wavefront because of the finite number of photons available for wavefront sensing . based on the analysis by guyon ( 2005 ) this allows one to derive fundamental limits for high - contrast imaging in the visible for a 30 m class telescope . we consider a perfect telescope with a perfect ao system and perfect coronagraphic instrumentation . we consider an ideal wavefront sensor making theoretically optimal use of all incoming photons ( a few existing wfs concepts do offer this level of sensitivity , but have not yet been deployed on telescopes ) . we only consider errors due to photon noise in the wavefront sensor , and assume that the ao system has no other source of error ( perfect dm , no calibration error ) . we assume that wfs is performed with the science detector , therefore removing any chromatic or non - common path issues . we also assume correction of both phase aberrations and amplitude ( scintillation ) by means of two perfect dms . for each spatial frequency , an optimal exposure time ( and therefore optimal control loop rate ) exists which optimizes the time lag effect on the corrected phase and the photon noise . better control , including for example predictive methods would improve these fundamental limits by a factor of a few but would not change the main conclusions . with these ideal assumptions , the residual aberrations from the atmosphere create a halo in the final raw image with the raw contrast given by @xmath69 here @xmath70 is the wind speed , f the star flux , @xmath71 the angular separation , @xmath72 is a term resulting from fresnel propagation between the turbulent layers , and the fried @xmath14 is defined at @xmath73 m ( see guyon 2005 for derivation and details ) . [ [ theoretical - limit ] ] theoretical limit + + + + + + + + + + + + + + + + + we assume a 30 m telescope with such a perfect system operating at the wavelength of the oxygen a - band ( 760 nm ) . the raw final image psf consists of a residual halo as shown in figure 4 . this does not correspond to the actual contrast sensitivity , since this residual halo can be subtracted further using calibration and differential methods . the ultimate sensitivity limit is set by the photon noise in this halo . as a template , we consider the case of the sun - earth system at 10pc , with star magnitude m=4.1 ( i band ) . we also assume 100% throughput and that a perfect noiseless halo can be subtracted . the required contrast for our template science case is @xmath1 at 0.1 arcsec . we find that the exposure time required to reach @xmath74 with a resolution r=70 is @xmath75s ( @xmath76 days ) . given the unrealistic set of assumptions adopted , we believe that this demonstrates that this observation is not feasible . [ [ technological - limitations ] ] technological limitations + + + + + + + + + + + + + + + + + + + + + + + + + there are numerous technical difficulties that will make it impractical to get close to the theoretical limit on an elt in visible light . this includes the required number of actuators , achievable loop rate , implementation of science camera wavefront sensing , coronagraph designs for segmented apertures , stability required for calibration and halo subtraction , etc . even at longer wavelengths , the @xmath1 contrast regime will be extremely challenging . a number of studies of exao instruments for 30 m class telescopes found that in practice one might reach @xmath77 contrast at @xmath78 mas and @xmath79 contrast at 100 mas in the near - ir ( @xmath80 m ) ( e.g. , macintosh et al . 2006 ; cavarroc et al . 2006 ; kasper et al . although this comes short of what is required for studying terrestrial planets , these instruments will have extremely interesting and exciting scientific capabilities for the broader study of exo - planetary systems . = 0.4 * figure 4 : * raw psf for a 30 m telescope observing at 760 nm with perfect telescope and exao . the psf is shown _ after _ correction with a perfect coronagraph , and is normalized to be @xmath81 at the center _ without _ coronagraph . it therefore shows the achieved contrast before application of any further differential methods . ideal wfs is assumed to be performed with the science detector at the same wavelength to eliminate non - common path errors or chromatic propagation effects . the turbulence parameters are standard . as discussed in the text , even such an idealized setup would not be sufficient for low - resolution spectroscopy of an earth - twin at 10pc . when combining fundamental limits and realistic technical capabilities for ao on the upcoming 30 m class telescopes , it is clear that space observations will be necessary for very high - contrast science at wavelengths shorter than @xmath2 m , as required for the study of terrestrial planets in the habitable zone of nearby stars . several space - based high - contrast projects are under study at this time ( e.g. , cash et al . 2009a , b ; kasdin et al . 2009 ; postman et al . the most exciting prospect is to address the habitability of terrestrial planets and search for life by identifying spectroscopic biomarkers ( o@xmath82 , o@xmath83 , h@xmath820 , ch@xmath84 , etc . ; kasting et al . adaptive optics has made exceptional advances in approaching space - like image quality at higher collecting area , although the exact prospects for optical - wavelength applications remain uncertain . this will provide exciting new access to scientific problems that were previously inaccessible . nonetheless , there will continue to be subjects that can _ only _ be studied at optical wavelengths from space , including wide - field imaging and ( medium - resolution ) spectroscopy at the deepest magnitudes , and the exceptional - contrast observations needed to characterize terrestrial exoplanets and search for biomarkers . so we expect that there will continue to be a strong complementarity between observations from the ground and from space in the coming decades . this provides strong motivating for continued technology development and new facilities , both for ground- and space - based optical ( and uv ) applications.=-2 [ astro 2010 abbreviations : noi : notice of intent ; sw : science frontiers white paper ; sp : state of the profession paper ] = 0.3truecm=1frogel , j. a. , et al . 2008 , `` aura s assessment of adaptive optics : present state and future prospects ( v4.0 ) '' ( http://www.aura-astronomy.org/nv/astro2010panel + docs / auras assessment of ao v4.pdf )

ground - based adaptive optics ( ao ) in the infrared has made exceptional advances in approaching space - like image quality at higher collecting area . optical - wavelength applications are now also growing in scope . we therefore provide here a comparison of the pros and cons of observational capabilities from the ground and from space at optical wavelengths . with an eye towards the future , we focus on the comparison of a @xmath0 m ground - based telescope with an 816 m space - based telescope . the parameters relevant for such a comparison include collecting area , diffraction limit , accessible wavelength range , background emission , atmospheric absorption and extinction , strehl ratio , field of view , temporal and spatial psf stability , and target accessibility in time and on the sky . we review the current state - of - the - art in ao , and summarize the expected future improvements in image quality , field of view , contrast , and low - wavelength cut - off . we compare the depth that can be reached for imaging and spectroscopy from the ground and from space in the v and j bands . we discuss the exciting advances in extreme ao for exoplanet studies and explore what the theoretical limitations in achievable contrast might be . our analysis shows that extreme ao techniques face both fundamental and technological hurdles to reach the contrast of @xmath1 necessary to study an earth - twin at 10 pc . based on our assessment of the current state - of - the - art , the future technology developments , and the inherent difficulty of observing through a turbulent atmosphere , we conclude that there will continue to be a strong complementarity between observations from the ground and from space at optical wavelengths in the coming decades . there will continue to be subjects that can _ only _ be studied from space , including imaging and ( medium - resolution ) spectroscopy at the deepest magnitudes , and the exceptional - contrast observations needed to characterize terrestrial exoplanets and search for biomarkers.additional contact information primary author : + matt mountain , stsci , [email protected] , 410 - 338 - 4710

observations show that all nearby galaxies with a massive bulge component host supermassive black holes ( smbhs ) ( e.g. , * ? ? ? * ; * ? ? ? smbhs accreting near the eddington limit ( @xmath4/@xmath5 @xmath0 0.11 ) are visible as luminous active galactic nuclei ( agn ) that often outshine their host galaxies . models of agn lifetime , constrained by observed eddington ratio distributions , suggest that smbh growth is dominated by this luminous phase , lasting @xmath0 a few @xmath6 10@xmath7 years ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? observations including constraints on the sizes of ionized `` bubbles '' around quasars ( e.g. , * ? ? ? * ; * ? ? ? * ) and the lengths of relativistic jets and radio lobes ( e.g. , * ? ? ? * ; * ? ? ? * ) suggest that the _ episodic _ lifetime of luminous activity is similar to the total lifetime , implying that a smbh is triggered to the luminous agn phase no more than a few times . smbhs therefore spend significant amounts of time in quiescent or low - activity phases , which may contribute up to @xmath020% of overall smbh growth @xcite . a smbh accreting at lower rates ( @xmath4/@xmath5 @xmath8 0.1 ) will appear as a low - luminosity agn ( llagn ) . llagn share several properties with more luminous agn , including similar spectral shapes ( e.g. , * ? ? ? * ) and similar radio - loud fractions and luminosity - dependent , intrinsic to optical flux ratios ( e.g. , * ? ? ? a more complete census of llagn is important for understanding smbh accretion history , but the relative significance of the host galaxy in llagn makes a full census of accretion activity a challenge . deep surveys have been effective at selecting a wide variety of agn , including luminous , unobscured agn as well as faint and/or obscured agn ( e.g. , see * ? ? ? * for a review ) . selection criteria usually include cuts on luminosity and spectral shape . multi - wavelength data further aid selection by allowing selection via , for example , excess emission compared to what is expected from optical flux ( e.g. , * ? ? ? * ) or radio luminosity ( e.g. , * ? ? ? the above methods have been successful in selecting a wide variety of agn , but nevertheless miss certain populations , such as very heavily obscured agn and llagn ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . while heavily obscured agn can often be selected in the ir ( e.g. , * * ; * ? ? ? * ; * ? ? ? * ) , the spectral energy distributions ( seds ) of llagn are likely dominated by the host galaxy in other bands . even in s , binaries , ultraluminous sources , and hot gas will provide significant contributions to the overall power output . observations of nearby low - luminosity seyfert nuclei artificially shifted to @xmath9 show that llagn would exhibit luminosities , spectral shapes , and -to - optical flux ratios consistent with those of normal or optically bright/ faint galaxies @xcite . by relying on such criteria , deep surveys may be underestimating agn fractions . variability potentially provides a useful indicator of whether an extragalactic source , classified as a galaxy by other means , harbors an agn . variability is a defining characteristic of agn and has long been used as an agn selection technique ( e.g. , * ? ? ? numerous studies have used optical variability to select agn from deep surveys such as the 1 ms , the subaru / xmm - newton deep field , and the goods north and south fields ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * respectively ) . spectroscopic observations of the 1 ms @xcite found that 17 of 27 optical variability - selected objects were broad - line agn ; 9 ( 5 ) agn would have been missed if selected by color ( selection ) . similarly , uv variability has been used successfully to identify llagn in galaxies with low - ionization nuclear emission - line regions ( liners ) . liners have been found in the nuclei of a large fraction of nearby galaxies ( e.g. , * ? ? ? * ; * ? ? ? * ) , but these regions could be ionized by either massive star clusters or low accretion - rate agn . _ hst _ imaging has found that @xmath025% of liners are associated with compact ( @xmath10 few pc ) uv sources @xcite . a study of liners with compact nuclear uv sources found significant variability in 15 of 17 , indicating the presence of an agn @xcite . deep x - ray surveys are able to detect variability in moderate - luminosity / high - redshift agn ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? the 4 ms _ chandra _ deep field - south @xcite , the deepest survey to date , allows a preliminary classification of agn on the basis of several observed quantities ( see [ sec : overview ] for details ) . this paper utilizes variability techniques to search for agn missed by these criteria . with 4 ms of exposure time spanning 10.8 years for 466 good - quality sources ( see [ sec : fluxvar ] ) , variability can be detected in sources with time - averaged fluxes as faint as @xmath11 ergs @xmath12 s@xmath13 . we use a cosmology with h@xmath14 = 70.4 km s@xmath13 mpc@xmath13 , @xmath15 = 0.272 , and @xmath16 = 0.728 ( e.g. , * ? ? ? the details of the 4 ms source catalog are available in @xcite ; we provide a brief summary here . the 4 ms , constructed from 54 _ chandra _ observations over 10.8 years , covers an area of 464.5 arcmin@xmath17 and reaches highest sensitivities of @xmath18 ergs @xmath12 s@xmath13 and @xmath19 ergs @xmath12 s@xmath13 , with multi - wavelength coverage in more than 40 bands from the radio to the uv . source candidates are detected using a 10@xmath20 false - positive probability threshold in wavdetect @xcite and are then pruned using a binomial no - source probability ( see appendix a2 of weisskopf et al . 2007 ) @xmath21 @xmath22 0.004 to obtain a more conservative list of 740 main - catalog sources , all of which are consistent with being point sources . source extraction and photometry were conducted with acis extract ( ae ; broos et al . ae models _ chandra _ s high resolution mirror assembly point spread function ( psf ) using the marx ray - tracing simulator . the psf model is used to generate a polygonal extraction region for each source that approximates the @xmath390% encircled energy fraction ( eef ) contour of a local psf measured at 1.497 kev . for @xmath226% of the candidates , the sources are crowded ( i.e. , the polygonal source regions overlap ) and smaller extraction regions that are as large as possible without overlapping ( @xmath23% eef ) are used . the background is calculated from regions that subtract the contribution from the source of interest and its neighboring sources ; the regions are typically a factor @xmath316 larger than the source - extraction region . ae merges the individual observations to estimate aperture - corrected , background - subtracted counts and the 1@xmath24 ( asymmmetric ) upper and lower statistical errors @xcite . in this paper , we will use the standard photometric bands : 0.52 kev ( soft ) , 28 kev ( hard ) , and 0.58 kev ( full ) . though most sources have a relatively small number of counts ( median net counts @xmath3 77 ) , a rough estimate of source spectral shape can be made by relating the band ratio ( i.e. , the ratio of the count rates in the 28 kev and 0.52 kev bands ) to an effective power - law photon index , @xmath25 ( @xmath26 @xmath27 @xmath28 @xmath29 @xmath30 ) . for low - count sources where @xmath25 can not be determined reliably , @xmath25 is set to 1.4 , the stacked ( co - added ) spectrum of all sources in the @xcite , which is consistent with the unresolved spectrum of the cosmic background @xcite . of 740 sources , 716 ( 96.8% ) contain matches in at least one of seven optical / near - infrared / radio ( onir ) catalogs : ( 1 ) the eso 2.2 m wfi @xmath31-band catalog @xcite , ( 2 ) the goods - s _ hubble space telescope _ ( @xmath32 ) version r2.0z @xmath33-band catalog @xcite , ( 3 ) the gems hst @xmath33-band catalog @xcite , ( 4 ) the goods - s music catalog @xcite , ( 5 ) the musyc @xmath34-band catalog @xcite , ( 6 ) the simple spitzer / irac 3.6@xmath35 m catalog @xcite , and ( 7 ) the vla 1.4 ghz radio catalog @xcite . of 716 sources with multi - wavelength identifications , 419 ( 58.5% ) have spectroscopic redshift measurements , collected from @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite . a total of 343 ( 81.9% ) of the 419 spectroscopic redshift measurements are `` secure '' , in that they are measured at @xmath36 95% confidence levels with multiple secure spectral features . 668 ( 93.3% ) sources have high - quality , accurate ( @xmath37/(@xmath38 ) @xmath3 6.5% ) photometric - redshift measurements from at least one of three photometric - redshift catalogs : @xcite , @xcite , and @xcite . the positions of primary onir counterparts were cross - matched with the photometric - redshift catalogs using a matching radius of 0.5@xmath39 , resulting in a false - match probability of @xmath101% . subsequent spectroscopic observations published in the arizona cdf - s environment survey ( aces ; cooper et al . 2012 ) catalog show with a blind test that errors on the photometric redshifts are @xmath101% . the 4 ms sources were classified as agn by the following criteria : + @xmath40@xmath41 high luminosities : @xmath42 @xmath43 ergs s@xmath13 , where the rest - frame luminosity has been corrected for galactic and intrinsic absorption . + @xmath40@xmath41 hard spectra : a source with @xmath25 @xmath22 1.0 is identified as a heavily obscured agn . + @xmath40@xmath41 high -to - optical flux ratios : log(@xmath44/@xmath45 ) @xmath46 , where @xmath44 = @xmath47 , @xmath48 , or @xmath49 and @xmath45 is the @xmath31-band flux . + @xmath40@xmath41 excess emission compared to that expected from star formation : @xmath50(@xmath51 @xmath52 ) @xcite , where @xmath52 is the 1.4 ghz monochromatic luminosity in w hz@xmath13 . + @xmath40@xmath41 an indication of broad emission lines in the optical spectrum . stars were classified by cross - matching sources ( using the onir counterpart positions ) with ( 1 ) the spectroscopically identified stars in @xcite , @xcite , and @xcite ; ( 2 ) the likely stars with stellarity indices greater than 0.7 in the gems _ hst _ catalog @xcite ; and ( 3 ) the likely stars with best - fit stellar templates in the musyc photometric - redshift catalog @xcite , using a matching radius of 0.5@xmath39 . sources not identified as an agn or a star were classified as galaxies . rest - frame 0.58 kev luminosities are calculated for all sources . for agn , which make up the vast majority of sources , the luminosity is corrected for galactic and intrinsic absorption . agn not detected in the full band have an upper limit on the luminosity based on the 3@xmath24 poisson error on the counts . for galaxies , the intrinsic - absorption correction and @xmath34-corrections may not be appropriate . of 92 galaxies with `` good quality '' observations ( see [ sec : fluxvar ] ) , 78 are not detected in the hard band and have poorly determined photon indices . all galaxies are detected in the soft - band . we perform a stacking analysis for galaxies with @xmath22 150 net counts ( 87 of 92 ) following the procedure described in [ sec : stack ] . the resulting average @xmath53 = [email protected] gives @xmath42/@xmath55 = 2.35 . we calculate the rest - frame 0.58 kev luminosity for galaxies as @xmath56 . since galaxy emission is typically unabsorbed , we do not apply any correction for intrinsic absorption . we perform two quality cuts before conducting variability analysis . first , we exclude the catalog sources with off - axis angles greater than 8@xmath57 to ensure that sources will have sufficient coverage ( @xmath5850 of 54 observations ) throughout the various pointings . to ensure accurate variability measurements , we also require that each source has at least 20 net counts in the 0.58.0 kev band ( i.e. , at least 5 counts on average per epoch , as defined below ) . these quality cuts result in a total sample of 466 sources : 369 classified as agn , 92 classified as galaxies , and 5 classified as stars in the @xcite catalog . all of these agn and galaxies have measured spectroscopic or photometric redshifts . the 92 sources classified as galaxies may nevertheless contain an agn , as would be indicated by variability , and make up the sample investigated in this paper . we divide the observations into four epochs , each containing @xmath01 ms of integration : 2000 ( 943.1 ks ) , 2007 ( 967.7 ks ) , 2010a ( march may ; 1015.5 ks ) , and 2010b ( may july ; 944.9 ks ) . as in @xcite , we merged observations within each epoch and , for a given source position from the catalog , measured the source and background counts and flux over three observed - frame energy bands : 0.58 kev , 0.52 kev , and 28 kev . a source is considered variable if the variability observed between observations is greater than that expected from poisson statistics , with a probability threshold of 5% that the variability is due to noise alone . ( the choice of probability threshold is discussed further below . ) to check whether a source can be considered variable , we calculate the quantity : + @xmath59 where @xmath60 is the number of epochs , @xmath61 is the photon flux ( background - subtracted counts with units of @xmath12 s@xmath13 ) in a given epoch , @xmath35 is the mean photon flux over all epochs , and @xmath62 is the error squared on the photon flux for the @xmath63 epoch . the photon flux is calculated by dividing the full - band ( 0.58.0 kev ) net counts by the exposure time and the mean effective area across the source aperture . the @xcite approximation gives the error on the net counts , which is propagated to obtain the error on the photon flux . since this error is significantly asymmetric for low - count ( @xmath64 counts ) sources , we average the upward and downward error bars for these objects to obtain the average error @xmath65 . ( the same method is applied in the monte carlo simulations below . ) for large photon fluxes , the @xmath66 statistic follows a @xmath67 distribution , and any source with @xmath66 @xmath58 7.82 ( for 3 degrees of freedom ) has a probability @xmath68 @xmath22 0.05 ( i.e. , 95% confidence level ) that the variability is due to random noise . however , at low count rates , the error on the photon flux is not gaussian . since errors in the low - count regime are larger than expected from a gaussian distribution , the resulting @xmath66 statistic is smaller than expected and does not follow the @xmath67 distribution ( see fig . 2 of @xcite ) . we therefore constructed a monte carlo simulation to determine the distribution the @xmath66 statistic should follow for each source , similar to the procedure followed by @xcite . we first scaled the total observed source and background counts for each source , obtained from the full 4 ms observation , to the exposure time and effective area for a given epoch . this procedure generates the source and background counts expected in each epoch if the source and background were constant over time , and it accounts for fluctuations in the background that will affect low - count sources . to simulate the variance expected from noise , poisson distributions were defined using the expected source and background counts as the mean values . we then simulated 1,000 observations of each source by repeatedly drawing the expected counts from the poisson distributions . for each simulated observation , we calculated the photon flux for four epochs and calculated @xmath66 as defined above . asymmetric errors on the source and background counts are obtained from @xcite and are propagated to get the error on the photon flux . the observed @xmath66 can then be compared to the simulated distribution to determine the probability @xmath68 that the observed variability is due to poisson noise . spurious sources of variability are negligible , since effective exposure maps are calculated separately for each observation , taking into account issues such as vignetting , ccd gaps , bad pixels , bad columns , and _ s spatial- and time - dependent quantum efficiency degradation . while using a more conservative @xmath69 = 1% on our dataset would result in fewer false positives , it would also eliminate a similar number of truly varying sources . for example , in the sample of 92 galaxies , @xmath69 = 5% results in 20 variable sources ( see [ sec : var ] ) , of which 4.6 are expected to be false positives . reducing @xmath69 to 1% results in 13 variable sources , of which 0.9 is expected to be a false positive . so while the more conservative critical value eliminates @xmath34 false positives , it also eliminates @xmath70 truly varying sources . the @xmath68 values listed in table 2 can be used to screen the sources further as desired . of the 369 sources classified as agn that meet both the off - axis angle ( @xmath71 @xmath22 8@xmath57 ) and count ( total net counts @xmath58 20 ) requirements , 50.1% exhibit significant flux variability ( @xmath68 @xmath22 0.05 ) on @xmath0month year timescales . for the 178 agn with more than 100 counts , 74.2% exhibit significant variability . the basic diagnostic plot in figure [ fig : fracvar ] shows the rise in the agn variable fraction with total net counts . the plot demonstrates that , given sufficient counts to detect it , variability is a near - ubiquitous property of faint , selected agn , even in the case of significant obscuration : @xmath370% of the sample consists of obscured agn @xcite , and 47.5% ( 51.5% ) of obscured ( unobscured ) agn are significantly variable . the variable fractions are consistent with the results from the 1 ms @xcite , although obtained with different temporal sampling and down to much fainter fluxes . in sources classified as galaxies , the variable fraction is significant at low counts and equals that of agn at higher counts ( figure [ fig : fracvar ] ) . table 1 describes the columns of table 2 , which lists the attributes of the 20 variable and 72 non - variable galaxies . ( variability properties of cdf - s agn will be covered in the forthcoming paolillo et al . , in preparation . ) the luminosity - redshift distribution of variable and non - variable galaxies is shown in figure [ fig : lz ] ; almost all lie below @xmath72 . spectroscopic redshifts are available for 18 of 20 variable galaxies and for 61 of 72 non - variable galaxies . photometric redshifts are available for the remaining sources . six example light curves ( background - subtracted count rates in the observed - frame 0.5 - 8 kev band vs. mjd ) representative of the sample as a whole are shown in figure [ fig : lc ] . the luminosity distributions of all galaxies and those exhibiting significant variability are shown in figure [ fig : lx_gal_histo ] . a k - s test shows that the two samples are consistent with being drawn from the same parent population ( @xmath73 = 57% ) , and the variable fraction does not show a significant dependence on luminosity below @xmath42 @xmath0 10@xmath74 ergs s@xmath13 . the possibility that agn - related variability may go undetected in galaxies is discussed in [ sec : nxsvar ] . we briefly compare agn selection based on variability to the following selection methods : ( 1 ) luminosity cuts , ( 2 ) the -to - optical flux ratio , and ( 3 ) excess emission compared to that expected from star formation , based on the radio luminosity . figure [ fig : fracvar_lx ] shows the fraction of variable sources vs. luminosity . below @xmath42 = 10@xmath75 ergs s@xmath13 , a luminosity cut often used for agn selection in surveys , the variable fraction remains significant at @xmath76 . of 64 galaxies with @xmath42 @xmath22 10@xmath75 ergs s@xmath13 , 17 ( @xmath327% ) are variable . agn selection via the -to - optical flux ratio is demonstrated by the @xmath31-band magnitude vs. flux plane in figure [ fig : xue ] ( cf . figure 16 in xue et al . sources classified as agn , galaxies , and stars in @xcite are marked as small red circles , larger black circles and blue stars , respectively . variable sources are marked with filled symbols . the optically bright , faint region , typically thought to exclude agn ( obxf ; @xmath44/@xmath77 ; hornschemeier et al . 2003 ) contains 27 galaxies . of these , 6 ( 22% ) are variable , with luminosities spanning log @xmath42 @xmath3 39.741.4 . agn may also be selected based on a comparison between and radio luminosities @xcite . the radio luminosity can be used to predict the luminosity in star - forming galaxies ( e.g. , * ? ? ? * and references therein ) , so sources with excess emission may be classified as agn . of the 17 galaxies with radio detections ( none of which have excess x - ray emission ) , three ( 18% ) are variable . these objects may have excess radio emission due to strong radio cores rather than star formation . due to the generally limited photon statistics of the galaxy sample , most variable sources must be strongly variable to be detected . significantly variable galaxies show maximum - to - minimum flux ratios @xmath78 @xmath3 1.59.3 ( median = 4.1 ) over the observed 10.8-year time frame . the smallest max - to - min ratio ( 1.5 ) was measured for [email protected] ( xid = 616 ) , a source with @xmath36500 counts . for most galaxies , total net counts are too low ( @xmath80 ) to detect variability below a factor of 23 . to address whether the variable galaxy max - to - min ratios sample the average agn population or only the highly variable `` tip of the iceberg , '' we ran a monte carlo simulation . the procedure assumed that the entire galaxy population is significantly variable and simulated the variability expected over 10.8 years of observation . following the procedure in @xcite , we used the @xcite algorithm to simulate 5,000 light curves based on the mean and variance of the flux for each of the 92 galaxies . for non - variable galaxies , the measured variance represents noise in the measured light curve , which gives an upper limit to the variability that could be present . the algorithm produces a random , continuously sampled light curve from a given power spectral density ( psd ) function , which we assumed to be described by a broken power - law , where the break frequency depends on smbh mass and accretion rate @xcite . since the break frequency lies outside the range of timescales sampled for most variable galaxies ( see [ sec : model ] ) , we simplify the model to a power - law ( @xmath21(@xmath81 ) @xmath27 @xmath82 ) with index @xmath83 , as is typical for the low - frequency ( long - timescale ) psds for nearby seyferts ( e.g. , * ? ? ? the time units of the light curves are determined by the minimum and maximum timescales input into the simulation ; the light curve durations are adjusted according to each source s redshift . we resampled / rebinned the light curve using the observing pattern and add poisson noise to the simulated light curve to account for measurement error . the full simulated light curves was made five times longer than the sampled region in order to produce variation on timescales much longer than those sampled by the data . this reproduces the effect that very long - timescale ( low - frequency ) variations have on variability measured over shorter timescales ( i.e. , `` red noise leak '' ) . we compare the simulated distribution of median max - to - min ratios , where the median is calculated over 5,000 trials for each galaxy , to the observed distribution for variable galaxies in figure [ fig : peaktopeak ] . a k - s test shows that these two populations have a 0.2% chance of being drawn from the same parent population . note that the simulated distribution ( black histogram ) illustrates an _ upper limit _ to the variability that could be present in the , resulting in a _ lower limit _ on the detectable fraction of sources . the detects at least @xmath318% of sources with max - to - min ratios @xmath58 2 and at least @xmath357% of sources with max - to - min ratios @xmath58 4 . a significant fraction of non - variable galaxies may still host an agn , but the variability may remain undetected due to low counts . for the galaxies exhibiting significant variability , we calculated the normalized excess variance ( e.g. , * ? ? ? * ) , which measures how strongly a source varies in excess of the measurement error . the excess variance is the integral of a source s psd function over a given frequency range , which is defined by the light curve s duration ( 10.8 years , observed - frame ) and minimum bin size ( 4.0 ms,1 ms in integration time , the timescale sampled is significantly longer due to the spread in _ observations , of which the shortest is 4.0 ms for the 2007 epoch @xcite . ] observed - frame ) . + @xmath84 where @xmath85 is the average of the asymmetric upward and downward measurement errors . using the upward ( downward ) error would overestimate ( underestimate ) the errors . zero excess variance ( @xmath86 ) would indicate that the observed count fluctuation is entirely consistent with noise rather than due to intrinsic source variability ; due to statistical fluctuations , the excess variance may also be negative in this case . the @xmath87 values are listed in table 2 . note that the variability amplitude is calculated for _ observed - frame _ energy bands . the variable galaxies cover a redshift range @xmath88 , so the excess variance will be measured over @xmath89 kev at @xmath33 = 0 to @xmath90 kev at @xmath33 = 1 . variability strength is known to depend on energy in some nearby seyfert galaxies ( e.g. , ark 120 , mcg6 - 30 - 15 , and i zw 1 ; vaughan et al . 2004 , vaughan & fabian 2004 , and gallo et al . 2007 , respectively ) , while in others , variability remains nearly constant with energy ( e.g. , ton s180 and nab 0205 + 024 ; vaughan et al . 2002 and gallo et al . 2004 , respectively ) . in the former cases , the change in variability strength is small , with a @xmath22 10% difference from 0.5 to 10 kev , so the bandpass effects at redshifts @xmath91 should remain small . we calculate the statistical error ( i.e. , measurement error ) on the excess variance following equation 11 of @xcite . @xmath92 the large errors on @xmath93 ( table 2 ) are due to the small numbers of counts observed for most sources ( e.g. , 12 of 20 variable galaxies have @xmath10 100 net counts ) ; four variable galaxies with net counts @xmath10 50 have excess variance measurements completely dominated by statistical uncertainty [ @xmath93 @xmath10 err(@xmath93 ) ] . nevertheless , most variable galaxies have excess variance measured at the @xmath361@xmath24 level . the excess variance contains additional sources of uncertainty aside from statistical error : ( 1 ) random scatter intrinsic to the stochastic nature of agn variability @xcite and ( 2 ) uncertainty and systematic bias due to sparse sampling of the light curve . the sparse pattern of observations will lead to large uncertainty in the mean flux measurement , and since the measured mean will be closer to the sampled data points rather than the true mean , the variance measurements will tend to be underestimated @xcite . we again employ a monte carlo simulation to model the uncertainty and systematic effects produced by random scatter and sparse sampling . we follow the same procedure described above to produce 5,000 light curves for every variable source . the mean and variance are calculated after sampling each simulated light curve with the observing pattern . the sampling bias can be corrected by rescaling the observed variance by a factor equal to the ratio between the `` true '' input variance ( i.e. , the observed variance used as input in the simulations ) and the median output variance ( i.e. , the biased variance affected by sparse sampling ) : @xmath94 = @xmath95/median(@xmath96 ) . the median output variance is calculated over all 5,000 light curves . the amount of systematic bias depends on the frequency and regularity of the sampling . if the sampling is regular , the scaling factor will approach unity as the number of samples increases ; however , the scaling factor will remain above unity even at high sampling frequency if the sampling is irregular @xcite . the slope of the psd will also affect the sampling bias a steeper psd slope ( i.e. , @xmath97 instead of @xmath83 ) will result in a larger bias for a given sampling pattern . since the intrinsic psd slopes are not known , we apply uniform corrections assuming @xmath83 ; however , source - to - source variations in psd slope may be a significant source of scatter in variability measurements . we find that the scaling factors range from @xmath94 @xmath3 0.03 to 4.8 , with a mean @xmath98 @xmath3 1.54 and a scatter on @xmath98 of @xmath99 . in @xmath376% of the sources , the scaling factor is greater than unity , indicating that the variance is underestimated due to the sampling bias . note that the monte carlo psd is normalized by each source s variance , which is calculated using the source s light curve . since the heavy binning could smear out high frequency variations , this method could result in an artificially smaller scaling factor . however , allevato et al . , in preparation , find a similarly small bias ( @xmath100 ) when sampling higher frequencies for a wide range of s / n ratios , gap lengths and sampling patterns . more importantly , the scatter on the @xmath94 factor calculated for each source is large ( @xmath101 ) , so individual measurements , even when corrected for bias , are likely to be poor estimates of the intrinsic variance . therefore , variability properties of galaxies are best considered in ensemble rather than on an individual basis . bias - corrected excess variances ( @xmath102 ) are listed in table 2 and are used for all further analysis unless otherwise noted . the three most likely sources of x - ray variability are binary ( xrb ) populations , ultraluminous sources ( ulxs ) , and accreting smbhs . in this section , we show that the first two possibilities are not likely to dominate the measured galaxy variability . to examine the potential contribution of binary populations to variability , we must first determine the relative contributions of low - mass binaries ( lmxb ) and high - mass binaries ( hmxb ) to the hard ( 210 kev ) , galaxy - wide luminosity ( @xmath103 ) . galaxy stellar mass ( @xmath104 ) scales the contribution of ( older ) lmxbs , and star formation rate ( sfr ) scales the contribution of ( younger ) hmxbs ( e.g. , equation 3 of * ? ? ? @xmath104 and sfr are calculated for each galaxy in @xcite using the optical colors and the uv and ir luminosities . both the @xcite relations and the @xcite calculations adopt the same initial stellar mass function @xcite . however , since @xcite and @xcite use different formalisms for computing stellar masses ( bell et al . 2003 and zibetti et al . 2009 , respectively , which differ primarily due to their models of star formation history ) , we apply a correction factor of 2.6 to the stellar masses from @xcite . the sfr and @xmath104 values for each variable galaxy are listed in table 2 . by comparing the expected lmxb and hmxb contributions to the total luminosity , we can determine which population ought to dominate the variability . in the variable galaxies , sfr ranges from 0.04 to 55 @xmath105 yr@xmath13 with a median of 2.2 @xmath105 yr@xmath13 ; @xmath104 ranges from @xmath106 to @xmath107 with a median of @xmath108 . we find that all but six of 20 variable galaxies are expected to have a larger hmxb contribution . @xmath109(sfr)/@xmath110(@xmath104 ) ranges from 0.01 to 75.8 with a median of 2.6 ( table 2 ) . hmxbs are generally more variable than lmxbs @xcite , so unless lmxbs dominate the output of a galaxy , we neglect their contribution . to determine the variability expected from the hmxbs , we follow the relations in 4.2.3 of @xcite , where the variability of the hmxb population is roughly determined by the galaxy s sfr . the following relations were obtained from monte carlo simulations with a power - law hmxb luminosity function with slope @xmath111 and a cut - off luminosity at @xmath112 = 2@xmath610@xmath113 ergs s@xmath13 : @xmath114 @xmath115 @xmath116 here , @xmath117 is the fractional rms ( i.e. , the square root of the excess variance ) expected from an individual binary , which can be as large as 2030% on @xmath0year timescales ( e.g. , * ? ? ? * ) , and @xmath118 is the total variability . we take @xmath119 and calculate the upper limit on @xmath118 . in the most extreme case , we find that the upper limit on variability expected from an hmxb population is @xmath120 . in six galaxies , lmxbs are expected to dominate the luminosity . four of the six have log @xmath104 @xmath58 10.5 @xmath121 , and therefore have large enough stellar mass to follow a @xmath122 law ( 4.4.2 of gilfanov et al . 2004 ) . for the remaining two galaxies , we follow the trend in figure 12 of @xcite . the upper limit on variability expected from an lmxb population is @xmath123 . we find that xrb populations can not explain the full extent of the variability for variable galaxies . figure [ fig : factor_var ] plots the distribution of @xmath102 for galaxies , showing that all the variable galaxies exhibit variability in excess of that expected from xrbs . the variable galaxies have a median @xmath102/@xmath124 @xmath3 42 ( without the bias correction , @xmath93/@xmath125 ) , indicating that the contribution of an xrb population to the measured variability is small . we also consider whether one or more ultraluminous sources ( ulxs ) may dominate a galaxy s output . the nature of ulxs is still debated , but most likely involves accretion onto massive stellar black holes ( 30100 @xmath105 ) ; a few cases may involve accretion onto intermediate mass black holes ( 100300 @xmath105 ) or beamed emission from 1020 @xmath105 black holes . typical luminosities span @xmath42 @xmath3 10@xmath12610@xmath127 ergs s@xmath13 ( e.g. , * ? ? ? * ; * ? ? ? in a survey of 1,441 point sources in 32 nearby galaxies in the _ chandra _ archive , @xcite found ulxs in 19 galaxies ; the contribution of one or more ulxs to a galaxy s total point source luminosity ranged from 7%87% , with a median contribution of 43% . ulxs could potentially explain nine variable galaxies with @xmath128 ergs s@xmath13 . however , since ulxs tend to be associated with star - forming regions ( e.g. , * ? ? ? * ) and occur more frequently in late - type / irregular galaxies than in early - type galaxies ( e.g. , * ? ? ? * ) , the five variable galaxies with both @xmath128 ergs s@xmath13 and late - type morphology ( see [ sec : morph ] ) are more likely to host ulxs . one method of finding ulxs is to search for off - nuclear sources . we plot postage - stamp images ( 8@xmath39 @xmath6 8@xmath39 ) of the goods - s / gems _ hst _ @xmath129-band for the variable galaxies in figure [ fig : stamp_var ] . the circle overplotted on each image has a radius 1.5 times the _ chandra _ positional erro , which is calculated at the 90% significance level . as in @xcite , sources offset from the galaxy nucleus by more than 1.5 times the positional error are considered off - nuclear . we find one marginally off - nuclear source : [email protected] ( xid = 269 ) . the primary optical source appears to be an early - type galaxy with blending toward the galaxy to the lower right , suggesting a merger . both galaxies have similar redshifts ( @xmath130 and 0.956 , respectively ) ; the first redshift is spectroscopic and secure ( see [ sec : overview ] ) and the second is photometric @xcite . the high luminosity ( @xmath42 @xmath3 2.2@xmath610@xmath75 ergs s@xmath13 ) suggests that a ulx is likely not the dominant source of emission from this galaxy . one variable galaxy , [email protected] ( xid = 418 ) , was previously identified as being off - nuclear in the 1 ms @xcite . with the additional data from the 4 ms , the x - ray source position has been refined ( with reduced uncertainty ) to be consistent with the galaxy s nucleus . the off - nuclear source discussed in the previous paragraph ( xid = 269 ) was not detected in the 1 ms , so it was not considered by @xcite . we can not rule out that a ulx may lie too close to a galaxy s nucleus to be detected as an off - nuclear source . however , the possibility of ulxs in most variable galaxies is mitigated by high luminosities and/or early - type morphology , so since we find no plausible off - nuclear ulxs , we conclude that ulxs are unlikely to dominate the emission from variable galaxies . accretion onto a smbh remains the best explanation of variable galaxies . we investigate the x - ray spectral shapes , the morphologies , and the optical spectral classifications of the variable galaxies for two purposes : ( 1 ) characterizing the variable galaxy population , and ( 2 ) determining whether their properties are consistent with those of llagn . as discussed in [ sec : overview ] , @xcite calculate the effective photon index for each source based on the ratio of count rates in the 28 kev and 0.52 kev bands ( table 2 ) . for 10 variable galaxies detected in the soft band but not the hard band , lower limits are listed . for 4 low - count variable galaxies detected either in the full - band , soft - band , or both , no reliable effective photon index can be determined , so @xmath25 is set to 1.40 . since 14 sources have poorly determined @xmath25 , we perform a stacking analysis on all variable galaxies with @xmath22 150 net counts to determine an average photon index for the sample . the three highest count sources ( with 199.2 , 275.6 , and 608.9 net counts ) are excluded from the stacking analysis since they could dominate the stacked signal , but the results remain the same within errors if these sources are included . following the procedure in @xcite , we calculate the soft and hard - band counts in a 3@xmath39 diameter aperture for each source . the background is calculated by averaging the counts in 1,000 randomly placed , source - free apertures within a 1@xmath57-radius circle around the source position . the individual source counts are summed and background is subtracted . aperture corrections are applied , averaged over all the observations weighted by exposure time ( see luo et al . 2011 for details ) , before calculating the band ratio . the stacked effective photon index for the 17 relevant variable galaxies is @xmath53 @xmath3 [email protected] , which is consistent with the typical photon index for local seyfert galaxies ( @xmath131 @xmath0 1.8 ; e.g. , dadina et al . 2008 ) at the 1@xmath24 level . including all 20 variable galaxies , @xmath53 @xmath3 [email protected] . while absorption may still be present in some individual sources , the soft spectrum implied by the stacked effective photon index indicates that the average variable galaxy is not heavily obscured . the x - ray luminosities are therefore intrinsically low and indicate that variable galaxies are most likely llagn . postage - stamp images ( 8@xmath132 ) of the goods - s / gems _ hst _ @xmath129-band for variable and non - variable galaxies are shown in figures [ fig : stamp_var ] and [ fig : stamp_non ] , respectively , with _ error circles overlaid . galaxies are classified by eye as late - type , early - type , irregular , or undetermined . mergers are also visually classified based on blending and/or tidal tails between two or more galaxies . since visual classifications are subjective and are particularly unreliable for distant , poorly resolved galaxies , we also apply the color - magnitude relation given in @xcite , where galaxies are considered part of the `` red sequence '' ( i.e. , early - type morphology ) if they are redder than ( @xmath133 @xmath134 ) = -0.31@xmath33 0.08@xmath134 0.51 . galaxies with bluer colors are considered part of the blue cloud ( i.e. , late - type morphology ) . although the color - magnitude diagram leads to a more objective classification scheme , there are nevertheless significant uncertainties , both in the rest - frame magnitudes and colors , and in the definition of the @xcite relation . moreover , the color classification is complicated by the fact that many of the galaxies lie in between the red sequence and blue cloud , in the so - called `` green valley . '' in cases where the color and the visually classified morphology disagree , we choose a final classification , preferring the visually classified morphologies in nearby , well - resolved galaxies , and the color classification in distant and/or poorly resolved galaxies . considering only galaxies classified as early or late - type , the visual and color classifications agree @xmath354% of the time . table 2 lists each galaxy s morphological type as determined by visual classification , the rest - frame magnitude and color ( from * ? ? ? * ) , the color classification according to @xcite , and the final classification . based on the final classification from table 2 , we find that variability does not prefer one morphology type over the other . variable galaxies have 40.0@xmath135% early - type morphology and 50.0@xmath136% late - type morphology , compared to 23.6@xmath137% and 51.4@xmath138% , respectively , in non - variable galaxies ( the errors are 1@xmath24 binomial errors ) . if we instead apply only the @xcite color classifications , then both the variable and non - variable galaxies prefer late - type morphologies ( 80.0@xmath139% and 72.2@xmath140% , respectively ) . the @xmath104 and sfr distributions for variable and non - variable galaxies show no significant difference ( @xmath73 @xmath58 10% ) . we find that the fraction of mergers among variable and non - variable galaxies is consistent within errors ( @xmath141% and @xmath142% , respectively ) . most ( 18/20 ) variable galaxies have optical spectral observations @xcite , from which spectroscopic redshifts were determined . in all cases , the optical spectra are classified as galaxies , showing only narrow emission lines or absorption lines . @xcite classify objects in more detail . of the eight variable galaxies listed in the @xcite catalog , two have typical galaxy spectra showing only absorption lines . the remaining six are classified as having low - excitation emission lines consistent with h ii region - type spectra . these objects would be classified as normal galaxies based on the optical data alone as the presence of the agn can not be established . however , one of these , [email protected] ( xid = 312 ) , has sufficient signal - to - noise in the optical spectrum to measure line ratios . this object is classified as a liner by @xcite via the line ratio diagnostics given by @xcite . in this section , we investigate how galaxy variability compares to agn variability . we first confirm a significant anti - correlation between excess variance and luminosity among agn , as seen in previous work ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the anti - correlation is plotted in figure [ fig : sigma_lx ] for significantly variable sources as @xmath102 vs. @xmath42 . the excess variance includes the bias correction discussed in [ sec : nxsvar ] . the rest - frame luminosities are calculated differently for agn and galaxies , as described in [ sec : overview ] . the spearman rank correlation coefficient ( @xmath143 = @xmath790.44 ) shows the correlation is significant at @xmath144 @xmath3 10@xmath145 ( 5.9@xmath24 ) for agn only ( red circles ) . the correlation increases in significance to @xmath146 ( @xmath147 ) if all variable sources above @xmath42 = 10@xmath127 ergs s@xmath13 are considered , including those classified as galaxies ( black circles and stars ) . no significant correlation is found if the sample is limited to only the variable galaxies , which are discussed further in [ sec : suppressed ] . note that the correlation coefficient and best - fit line parameters ( given below ) remain consistent within errors if calculated using the excess variance uncorrected for sampling bias ( @xmath93 ) . since both variability and luminosity may depend on other parameters , such as black hole mass and/or accretion rate , we fit the data using sixlin.pro , an idl program adapted from @xcite . a least - squares bisector fit to sources with @xmath148 @xmath58 10@xmath127 ergs s@xmath13 , weighted by the uncertainties in @xmath149 , results in @xmath149 = ( [email protected] ) @xmath150 . the slope and intercept are consistent within errors if the weights are not included . large blue squares in figure [ fig : sigma_lx ] show the weighted means for each luminosity bin ; as expected , these are consistent with the weighted least - squares bisector . note that since weighted means will most closely follow data with small error bars , they are weighted heavily downward in this log - log plot . limiting the sample further to sources with @xmath42 @xmath58 10@xmath75 ergs s@xmath13 results in @xmath149 = ( [email protected] ) @xmath151 . the slopes are both consistent with the results of @xcite , where a weighted least - squares bisector fit results in @xmath87 @xmath27 @xmath152 ( nandra et al . 1997 do not give the normalization of their relation ) . this result is notable because the @xcite data sampled shorter ( hour day ) timescales compared to the months years timescales sampled by the . the slope presented in this paper is significantly flatter than that found in paolillo et al . ( 2004 ; @xmath87 @xmath27 @xmath153 ) , which included non - varying sources . we choose not to include upper limits for non - varying sources in our analysis because the assumptions underlying survival analysis , which have been successfully applied to deal with censored data in other astronomical situations , do not apply here because : ( 1 ) the excess variance measurements of most sources lie near the detection limit , ( 2 ) a large percentage of sources do not have detected variability ( @xmath350% of agn and @xmath378% of galaxies ) , and ( 3 ) a significant percentage of sources with no detected variability , especially those classified as galaxies , may truly not be variable ( i.e. , @xmath93 @xmath0 0 ) . by not including censored data in this paper s measurements , we likely bias the measured slopes and possibly the significance of the anti - correlation . nevertheless , the variability - luminosity anti - correlation has been observed to follow a model based on smbh mass and accretion rate ( e.g. , * ? ? ? * ) , suggesting that the anti - correlation is real , though probably not linear . the model is discussed further in [ sec : model ] . we check for other potential biases that may affect the luminosity - variability anti - correlation . the log - log plot of figure [ fig : sigma_lx ] has the disadvantage of `` hiding '' sources with negative @xmath102 values . sources with low flux values , and hence higher scatter in @xmath102 will therefore appear to have stronger variability , since the @xmath102 values scattered to negative values will be hidden . to check for this bias , we remove high scatter sources with @xmath154(@xmath102 ) @xmath58 0.1 , excluding all but 38 sources , and find that the @xmath155@xmath42 correlation remains significant at @xmath144 @xmath3 0.9% ( 2.6@xmath24 ) . the best - fit slope and intercept remain consistent at the 1@xmath24 level . the flux - limited nature of the survey presents another potential bias . since luminosity correlates with redshift , and intrinsic variability timescales decrease with redshift , the decrease in variability strength as luminosity increases could , in principle , simply reflect the fact that shorter timescales are studied at higher redshifts , and therefore exhibit less variability due to the red - noise nature of agn light curves . to check for a possible redshift bias , we examine the @xmath102-@xmath42 correlation using a sub - sample within a narrow redshift range ( 0.55 @xmath22 @xmath33 @xmath22 0.75 ) . this redshift range selects 10% of the total sample , covers luminosities from @xmath42 @xmath0 10@xmath156 to @xmath157 ergs s@xmath13 , and results in minimal differences in rest - frame timescales . the spearman rank correlation coefficient for this sub - sample remains significant at @xmath144 @xmath3 0.8% ( 2.6@xmath24 ) ; the slope and intercept are consistent at the 1@xmath24 level . the above test also addresses another potential bias due to the redshift range of the sample . the @xmath158 values listed in table 2 measure the variability strength in the observed frame , so they will sample different energy bands depending on the source s redshift ( see [ sec : nxsvar ] for discussion ) . this could introduce bias if the variablity amplitude changes with increasing energy : from @xmath89 kev at @xmath33 = 0 to @xmath90 kev at @xmath33 = 1 . the slope and intercept remain the same within the narrow redshift range tested above , suggesting that any such bias does not have a significant effect on the sample . variable sources with luminosities less than @xmath42 = 10@xmath127 ergs s@xmath13 tend to fall significantly below the extrapolated linear relation by factors of @xmath3680 ( median factor of @xmath324 ) , indicating a drop in variability relative to the linear relation on long timescales for llagn . this suppressed " variability can be shown to be intrinsic to properties of agn variability rather than due to dilution by unrelated xrb populations . in [ sec : comparisons ] , we showed that galaxy variability can not be attributed solely to xrb populations . we now check whether the xrb contribution could nevertheless dilute the observed variability by estimating how much xrbs are expected to contribute to the total hard ( 210 kev ) , galaxy - wide luminosity ( @xmath103 ; see [ sec : comparisons ] ; lehmer et al . 2010 ) . we compare @xmath103 to the measured , intrinsic luminosity . first , @xmath42 is converted to @xmath159 , using the intrinsic photon index of @xmath131 = 1.8 adopted by @xcite for agn , and the stacked photon index @xmath53 @xmath3 1.9 for galaxies ( see [ sec : overview ] ) . figure [ fig : compare_lx ] plots the measured 210 kev luminosity against that expected from an xrb population for variable galaxies ( red stars ) and non - variable galaxies ( green squares ) . for reference , we plot in the same figure 369 agn ( open orange circles ) , 32 local galaxies @xcite , and 20 local luminous infrared galaxies , which are likely to be actively star - forming @xcite . the red solid line shows unity , while the dotted lines mark the dispersion observed in @xcite . a k - s test shows no significant difference between the relative xrb contribution in variable and non - variable galaxies . the galaxies on average tend to lie above unity , with a median @xmath159/@xmath103 @xmath3 4.6 . the excess in the non - variable galaxy population is perhaps not surprising given the flux - limited nature of the survey , which is more likely to detect the objects at the bright end of the galaxy luminosity function , and may result in a high percentage of contamination " by agn . the median excess for variable galaxies ( @xmath160/@xmath161 @xmath3 9.2 ) suggests that xrbs contribute @xmath011% of the 210 kev luminosity for the average variable galaxy . for six variable galaxies , the total x - ray emission is consistent with that expected from xrb emission within the scatter of the @xcite relation ( see fig . [ fig : compare_lx ] ) , suggesting that dilution may be more significant in these sources . three of these sources have variability consistent with the linear variability - luminosity relation , while three have suppressed variability ( filled black stars in fig . [ fig : sigma_lx ] ) . dilution by xrb variability may therefore play a role in suppressed variability , but can not fully explain the extent to which @xmath102 is suppressed at low luminosities . an alternative possible explanation for the suppressed variability at low luminosities is a change in accretion structure . @xcite found a similar drop in variability strength below @xmath159 @xmath3 2@xmath610@xmath127 ergs s@xmath13 in a sample of llagn and liners observed with _ asca _ on variability timescales of less than a day . the authors hypothesized that a radiatively inefficient accretion flow ( riaf , e.g. , yuan & narayan 2004 ) could be responsible for suppressed short - timescale variability at low luminosities due to the larger extent of the source . this scenario would not obviously explain the reduced variability on @xmath0month year timescales seen here . riaf models also predict a hard x - ray photon index due to the lack of an optically thick accretion disk , which provides the soft x - ray photons . the stacked x - ray photon index for variable galaxies ( @xmath53 @xmath3 [email protected] ; [ sec : stack ] ) is inconsistent with this prediction . studies since @xcite have found evidence both for `` suppressed '' variability in llagn ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) and against it ( e.g. , * ? ? ? * ; * ? ? ? similarly , objects such as narrow - line seyfert 1 ( nls1 ) galaxies and the dwarf seyfert ngc 4395 ( @xmath162 @xmath3 @xmath163 ) exhibit `` excess '' vairability for their luminosity @xcite . however , when plotting variability against mass instead of luminosity ( e.g. , * ? ? ? * ; * ? ? ? * ) , such discrepancies disappear , with residual differences possibly due to varying accretion rates ( e.g. , * ? ? * ; * ? ? ? * ) . to investigate the role of smbh mass ( @xmath162 ) and accretion rate ( normalized by the eddington rate ; @xmath164 = @xmath165/@xmath166 ) , we have obtained rough estimates for all variable sources . masses and eddington ratios for all galaxies , variable and non - variable , are listed in table 2 . the smbh masses are estimated via the scaling relation between @xmath162 and absolute @xmath34-band magnitude @xcite : @xmath167 the total absolute rest - frame magnitudes are derived from sed fitting in @xcite with a random scatter of @xmath100.3 mag . an luminosity - dependent correction factor ( equation 1 of * ? ? ? * ) corrects for nuclear emission . we assume that the host galaxy is bulge - dominated , a valid assumption for most agn ( e.g. , * ? ? ? * ; * ? ? ? several variable galaxies , however , are not bulge - dominated ( nine are late - type ; see table 2 ) , so their black hole masses may be overestimated . we apply a luminosity - dependent bolometric correction ( @xmath168 ) to estimate the bolometric luminosity @xcite and calculate the eddington ratio ( @xmath169/@xmath5 @xmath29 [ @xmath168 @xmath159]/[@xmath171 ( @xmath162/@xmath121 ) ] ) . note , that the @xcite correction is calculated explicitly for @xmath159 @xmath58 10@xmath75 ergs s@xmath13 ; we extrapolate this relation down to the lower luminosities of the variable galaxy sample . both the @xmath162 and @xmath169 estimation techniques are known to have large dispersions . the @xmath162-@xmath172 scaling relation has a total scatter of 0.33 dex , and additional uncertainty will come from the luminosity - dependent correction for nuclear emission , which is based on template seds @xcite . in addition , the assumption that all the variable galaxies are bulge - dominated will produce additional scatter . the bolometric correction , too , has large scatter due to the intrinsic dispersion in the sed . the uncertainty in @xmath169 due to sed dispersion is @xmath020% for luminous agn @xcite . there is some debate regarding the similarity between the seds of llagn and luminous agn ( e.g. , [ sec : intro ] ; ho 1999 and ho 2002 vs. maoz et al . 2007 ) ; however , the luminous agn sed dispersion likely serves as a lower limit to dispersion in llagn seds . variable galaxies tend to have lower accretion rates ( @xmath173 = 4 @xmath6 10@xmath174 ) and masses ( @xmath175@xmath176 ) than variable agn ( @xmath173 = 9 @xmath6 10@xmath177 ; @xmath175@xmath178 ) , where we have limited the agn sample to @xmath33 @xmath22 1 for purposes of comparison . a k - s test shows that the differences in the @xmath164 and @xmath162 distributions are significant : p@xmath179 @xmath3 1.7@xmath610@xmath20 and p@xmath179 @xmath3 0.002 , respectively . the properties of variable agn and galaxies are consistent with the range of estimates made by @xcite for selected , @xmath33 @xmath22 1 agn in the 1 ms , which span @xmath164 @xmath0 10@xmath180 ( median @xmath3 0.001 ) and @xmath181 ( median @xmath3 @xmath182 @xmath105 ) . unlike previous studies @xcite , we find no significant ( anti-)correlation between @xmath158 and @xmath162 , most likely due to the large scatter in @xmath158 and @xmath162 measurements , combined with the narrow range of masses probed ( since most sources lie between @xmath162 @xmath0 10@xmath18310@xmath184 @xmath105 ) . it is not surprising , therefore , that we find a significant anti - correlation between @xmath158 and @xmath164 ( @xmath185 , @xmath186 ) , plotted in figure [ fig : mdot ] , which is likely an artefact of the @xmath158@xmath42 anti - correlation . nevertheless , it is interesting that in figure [ fig : mdot ] the variable galaxies ( black circles ) connect smoothly with the more luminous agn ( red circles ) , with no discrepancy in variability strength . this suggests that the factor of 22.5 difference in @xmath164 may explain the `` suppressed '' variability in variable galaxies . a number of recent studies ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) have shown that variability may be determined by a combination of @xmath162 and @xmath164 , explaining the observed variability - luminosity relation . agn light curves appear to be universally described by a broken power - law psd function , where the break frequency depends on mass and accretion rate : @xmath187(@xmath162/@xmath188)@xmath13 , where @xmath189 is the accretion efficiency , assumed to be @xmath190 @xcite . the excess variance is equivalent to the integral of the psd between the minimum and maximum frequencies sampled by the data , so as long as the break frequency falls within this range , the excess variance at a given luminosity will decrease with increasing @xmath162 and increase with increasing @xmath164 . we compare the @xmath162 and @xmath164 values estimated in [ sec : suppressed ] to the variability measured for the variable agn and galaxy populations . assuming a universal psd function , we derive the variability - luminosity relations expected for the sampling pattern , given a range of smbh masses and accretion rates ( e.g. , * ? ? ? * and references therein ) . the bolometric luminosity , which is calculated directly from @xmath162 and @xmath164 , is converted to luminosity via a bolometric correction @xcite . where the break frequency lies outside the timescales sampled by the data ( depending on the combination of @xmath164 and @xmath162 ) , the excess variance will remain constant . the model variability - luminosity relations are plotted in figure [ fig : sigma_lx_physics ] for the average accretion rates for variable ( @xmath164 = 9 @xmath6 10@xmath177 ; dashed line ) and non - variable ( @xmath164 = 4 @xmath6 10@xmath174 ; dash - dotted line ) galaxies covering @xmath162 = 10@xmath19110@xmath192 @xmath105 ( low to high @xmath42 ) . both variable populations are limited to @xmath33 @xmath22 1 for comparison , and the relations are calculated at @xmath33 = 0.5 and @xmath33 = 0.7 , the median redshifts for variable galaxies and agn , respectively ; using model redshifts @xmath33 = 0 or 1 resulted in negligible changes . the shape of the observed variability - luminosity anti - correlation , including the plateau at low luminosity , is roughly reproduced by the model . the plateau occurs when the break frequency lies outside the range of timescales sampled by the data . unfortunately , the long timescales ( especially the long _ minimum _ timescale of 4 ms ; see [ sec : nxsvar ] ) and the large scatter prevent the models from distinguishing between the significantly different accretion rates estimated for variable galaxies and agn . note that most sources are more variable than predicted by the model . we note some possible sources of bias : ( 1 ) the normalization of the psd function used to calculate the models is based on a small sample of nearby agn @xcite and may therefore not be representative of agn out to @xmath193 . ( 2 ) both the models and the bias correction applied to the measured excess variance ( [ sec : nxsvar ] ) depend on a universal broken power - law psd function , but the slopes of the psd may vary between individual sources , and some sources may even have a second break at shorter frequencies ( e.g. , * ? ? ? ( 3 ) peculiar sources , such as nls1s , exhibit higher variability , so the variable sample may suffer from selection effects due to the higher sensitivity to more variable sources ( [ sec : nxsvar ] ) . we have investigated the variability properties of sources classified as galaxies in the 4 ms by dividing the observations into @xmath31 ms epochs . we find the following main results : \1 . using a @xmath67-based monte carlo simulation appropriate for low - count sources , 20 of 92 galaxies are selected as variable , as well as 185 of 369 agn . variability is effective in selecting agn that do not meet other agn selection criteria , such as cuts based on luminosity , -to - optical flux ratio , or excess emission compared to radio emission expected from star - forming galaxies ( see [ sec : fluxvar ] ) . if all 20 variable galaxies are accreting smbh , then the commonly used selection criteria employed by @xcite miss 20 of 389 agn ( @xmath35.1% ) , and 19 of 54 llagn ( @xmath335.2% ; where llagn are arbitrarily defined as having @xmath42 @xmath22 10@xmath75 ergs s@xmath13 ) . due to the low counts measured for most galaxies , it is likely that some non - variable galaxies may still host an agn but not exhibit detectable variability , so these fractions serve as lower limits . we measure variability strength with the normalized excess variance , accounting for measurement error , red - noise scatter , and bias due to the sampling pattern . comparing the bias - corrected excess variance to that expected from xrb populations , we find that xrbs can not explain galaxy variability . galaxy x - ray variability is most likely associated with accretion onto a relatively unobscured smbh . though some absorption may be present , the individual and stacked photon indices show no indication of heavy obscuration and are consistent with the typical agn spectral shape ( @xmath53 @xmath0 [email protected] ) . we confirm a significant anti - correlation between excess variance and luminosity , and find the slope and intercept to be consistent with what is reported in the literature for shorter timescales . we show that the sampling bias induced by the observing pattern does not affect the anti - correlation slope or significance . low - luminosity sources ( @xmath42 @xmath22 10@xmath127 ergs s@xmath13 , largely variable galaxies ) have `` suppressed '' variability compared to the extrapolated linear relation between excess variance and luminosity . this may be explained by their lower accretion rates . variable galaxies sample a different mass - accretion rate space than the rest of the agn population , with significantly lower accretion rates ( @xmath173 = 4 @xmath6 10@xmath174 ) and masses ( @xmath175@xmath176 ) than variable agn ( @xmath173 = 9 @xmath6 10@xmath177 ; @xmath175@xmath178 ) in the same redshift range ( @xmath194 ) . we find that an empirical model based on a universal broken power - law psd function , where the break frequency depends on smbh mass and accretion rate , roughly reproduces the shape of the variability - luminosity anti - correlation . however , the normalization of the model is low compared to the data , suggesting systematic bias either within the data ( e.g. , selection effects ) or the model ( e.g. , incorrect normalization for llagn at @xmath33 @xmath10 1 ) . we have shown that the variability measured by deep x - ray surveys is an effective technique for selecting cosmologically distant llagn . extending the to even longer exposures would enable detection of variability in both fainter and less variable sources , thus allowing better characterization of the properties and abundance of the llagn population . within the 4 ms , smaller time bins could be used to search for variability and better characterize the variability presented here . in addition , follow - up optical spectroscopy is necessary to characterize the llagn population presented here . while @xcite were able to classify one object as a liner , the spectra for the other sources did not have sufficient signal - to - noise to measure line ratios , so these objects would require either deeper optical spectroscopy or stacking of the currently available spectra . we thank p. uttley , e. feigelson , and c. saez for helpful discussions on variability statistics . we also thank m. gilfanov for helpful discussions regarding xrb variability . we thank the referee for constructive comments . we acknowledge the financial support of nasa adp grant nnx10ac99 g ( m.y . , w.n.b . , y.q.x . ) and chandra x - ray center grant sp1 - 12007a ( w.n.b . , y.q.x . ) . m.p . acknowledges support from prin-2009 by the italian miur . f.e.b . acknowledges support from programa de financiamiento basal and conicyt - chile under grants fondecyt 1101024 , alma - conicyt 31100004 , and fondap - cata 15010003 . b.d.l . acknowledges financial support provided by the einstein fellowship program . thanks the science and technology facilities council ( stfc ) for support . alexander , d. m. , bauer , f. e. , chapman , s. c. , smail , i. , blain , a.w . , brandt , w. n. , & ivison , r. j. 2005 , apj , 632 , 736 + alexander , d. m. et al . 2008 , apj , 687 , 835 + allevato , v. et al . 2011 , in preparation + almaini , o. et al . 2000 , mnras , 315 , 325 + babic , a. , miller , l. , jarvis , m. j. , turner , t. j. , alexander , d. m. & croom , s. m. 2007 , a&a , 474 , 755 + balestra , i. , et al . 2010 , a&a , 512 , 12 + barr , p. & mushotzky , r. f. 1986 , nature , 320 , 421 + barmby p. et al . 2006 , apj , 650 , l45 + barth , a. j. , ho , l. c. , filippenko , a. v. , & sargent , w. l. w. 1998 , apj , 496 , 133 + bauer , f. e. , alexander , d. m. , brandt , w. n. , schneider , d. p. , treister , e. , hornschemeier , a. e. , & garmire , g. p. 2004 , aj , 128 , 2048 + bell , e. f. , mcintosh , d. h. , katz , n. , & weinberg , m. d. 2003 , apjs , 149,289 + bell , e. f. , wolf , c. , meisenheimer , k. , rix , h .- w . , borch , a. , dye , s. , kleinheinrich , m. , wisotzki , l. & mcintosh , d. h. 2004 , apj , 608 , 752 + binder , b. , markowitz , a. & rothschild , r. e. 2009 , apj , 691 , 431 + blundell , k. m. , rawlings , s. , & willott , c. j. 1999 , aj , 117 , 677 + boller , t. , brandt , w. n. & fink , h. 1996 , a&a , 305 , 53 boutsia , k. , leibundgut , b. , trevese , d. & vagnetti , f. 2009 , a&a , 497,81 + brandt , w. n. & hasinger , g. 2005 , ara&a , 43 , 827 + broos , p. s. , townsley , l. k. , feigelson , e. d. , getman , k. v. , bauer , f. e. , & garmire , g. p. 2010 , apj , 714 , 1582 + caldwell , j. a. r. , et al . 2008 , apjs , 174 , 136 + cameron , e. 2011 , pasa , 28 , 128 + cardamone , c. n. , et al . 2010 , apjs , 189 , 270 + colbert , e. j. m. , heckman , t. m. , ptak , a. f. , strickland , d. k. & weaver , k. a. 2004 , apj , 602 , 231 + cooper , m. c. et al . 2012 , mnras , submitted , arxiv:1112.0312 + dadina , m. 2008 , a&a , 485 , 417 + damen , m. , et al . 2011 , apj , 727 , 1 + elvis , m. et al . 1994 , apjs , 95 , 1 + ferrarese , l. & ford , h. 2005 , space sci . rev . , 116 , 523 + freeman , p. e. , kashyap , v. , rosner , r. , & lamb , d. q. 2002 , apjs , 138 , 185 + gallo , l. c. , boller , th . , brandt , w. n. , fabian , a. c. & vaughan , s. 2004 , mnras , 355 , 330 + gallo , l. c. , brandt , w. n. , costantini , e. & fabian , a. c. 2007 , mnras , 377 , 1375 + gehrels , n. 1986 , apj , 303 , 336 + giavalisco , m. , et al . 2004 , apj , 600 , l93 + gilfanov , m. , grimm , h. j. & sunyaev , r. 2004 , mnras , 351 , 1365 + gilfanov , m. 2010 , lnp , 794 , 17 + goncalves , t. s. , steidel , c. c. , & pettini , m. 2008 , apj , 676 , 816 + gonzalez - martin , o. , papadakis , i. , reig , p. & zezas , a. 2011 , a&a , 526 , 132 + graham a. 2007 , mnras , 379 , 711 + grazian , a. , et al . 2006 , a&a , 449 , 951 + grogin n. a. et al . 2005 , apj , 627 , 97 + gultekin , k. , et al . 2009 , apj , 698 , 198 + hawkins , m. r. s. 2000 , a&a , 143 , 465 + hickox , r. c. & markevitch , m. 2006 , apj , 645 , 95 + hopkins , p. f. & hernquist , l. 2009 , apj , 698 , 1550 + hornschemeier , a. e. et al . 2003 , apj , 126 , 575 + ho , l. c. , filippenko , a. v. , & sargent , w. l. w. 1993 , apj , 417 , 63 + ho , l. c. , filippenko , a. v. , & sargent , w. l. w. 1997 , apjs , 112 , 315 + ho l. c. 1999 , apj , 516 , 672 + ho l. c. 2002 , apj , 564 , 120 + houck , j. r. et al . 2005 , apjl , 622 , 105 isobe , t. , feigelson , e. d. , akritas , m. g. & babu , g. j. 1990 , apj , 364 , 104 + iwasawa , k. , tanaka , y. & gallo , l. c. a&a , 514 , 58 + jakobsen , p. , jansen , r. a. , wagner , s. , & reimers , d. 2003 , a&a , 397,891 + kauffmann , g. , et al . 2003 , mnras , 346 , 1055 + kellermann , k. i. , sramek , r. , schmidt , m. , shaffer , d. b. & green , r. 1989 , aj , 98 , 1195 + kormendy j. & gebhardt k. , 2001 , in martel h. , wheeler j. c. , eds , the 20th texas symposium on relativistic astrophysics . am . phys . , new york , 586 , 363 + kroupa , p. 2001 , mnras , 322 , 231 + komatsu , e. , et al . 2011 , apjs , 192 , 18 + lawrence , a. & papadakis , i. 1993 , apj , 414 , 85l + le fevre , o. , et al . 2004 , a&a , 428 , 1043 lehmer , b. d. , brandt , w. n. , hornschemeier , a. e. , alexander , d. m. , bauer , f. e. , koekemoer , a. m. , schneider , d. p. & steffen , a. t. 2006 , aj , 131 , 2394 + lehmer , b. d. , et al . 2008 , apj , 681 , 1163 + lehmer , b. d. , alexander , d. m. , bauer , f. e. , brandt , w. n. , goulding , a. d. , jenkins , l. p. , ptak , a. & roberts , t. p. 2010 , apj , 724 , 559 + luo , b. , et al . 2008 , apjs , 179 , 19 + luo , b. , et al . 2010 , apjs , 187 , 560 + luo , b. , et al . 2011 , apj , 740 , 37 + marconi , m. , risaliti , g. , gilli , r. , hunt , l. k. , maiolino , r. & salvati , m. 2004 , mnras , 351 , 169 + markowitz a. & uttley p. 2005 , apj , 625 , l39 mchardy , i. m. , papadakis , i. e. , uttley , p. , page , m. j. & mason , k. o. 2004 , mnras , 348 , 783 + mchardy , i. m. , koerding , e. , knigge , c. , uttley , p. & fender , r. p. 2006 , nature , 444 , 730 + mchardy , i. m. , arvalo , p. , uttley , p. , papadakis , i. e. , summons , d. p. , brinkmann , w. & page , m. j. 2007 , mnras , 382 , 985 + maoz , d. 2007 , mnras , 377 , 1696 + maoz , d. , nagar , n. m. , falcke , h. & wilson , a. s. 2005 , apj , 625 , 699 + maoz , d. , filippenko , a.v . , ho , l.c . , rix , h .- w . , bahcall , j. n. , schneider , d . p. , & macchetto , f. d. 1995 , apj , 440 , 91 + martinez - sansigre , a. , rawlings , s. , lacy , m. , fadda , d. , marleau , f. , simpson , c. , willott , c. j. , & jarvis , m. j. 2005 , nature , 436 , 666 + mignoli , m. , et al . 2005 , a&a , 437 , 883 + miller , n. a. , fomalont , e. b. , kellermann , k. i. , mainieri , v. , norman , c. , padovani , p. , rosati , p. , & tozzi , p. 2008 , apjs , 179 , 114 + miniutti , g. , ponti , g. , greene , j. e. , ho , l. c. , fabian , a. c. & iwasawa , k. 2009 , mnras , 394 , 443 morokuma , t. 2008 , apj , 676 , 121 + nandra , k. , george , i. m. , mushotzky , r. f. , turner , t. j. & yaqoob , t. 1997 , apj , 476 , 70 + oneill , paul m. , nandra , k. , papadakis , i. e. & turner , t. j. 2005 , mnras , 358 , 1405 paolillo , m. , schreier , e. j. , giacconi , r. , koekemoer , a. m. , & grogin , n. a. 2004 , apj , 611 , 93 + papadakis i. e. 2004 , mnras , 348 , 207 + papadakis , i. e. , ioannou , z. , brinkmann , w. & xilouris , e. m. 2008a , a&a , 490 , 995 + papadakis , i. e. , chatzopoulos , e. , athanasiadis , d. , markowitz , a. & georgantopoulos , i. 2008b , a&a , 487 , 475 + peterson , k. c. , gallagher , s. c. , hornschemeier , a. e. , muno , m. p. & bullard , e. c. 2006 , aj , 131 , 133 + pian , e. , romano , p. , maoz , d. , cucchiara , a. , pagani , c. & la parola , v. 2010 , mnras , 401 , 677 + popesso , p. , et al . 2009 , a&a , 494 , 443 + ptak , a. , yaqoob , t. , mushotzky , r. , serlemitsos , p. & griffiths , r. 1998 , apj , 501 , 37l + ptak , a. , terashima , y. , ho , l. c. & quataert , e. 2004 , apj , 606 , 173 + rafferty , d. a. , brandt , w. n. , alexander , d. m. , xue , y. q. , bauer , f. e. , lehmer , b. d. , luo , b. , & papovich , c. 2011 , apj , 742 , 3 + ravikumar , c. d. , et al . 2007 , a&a , 465 , 1099 + richards , g.t . 2006 , apjs , 166 , 470 sarajedini , v. , koo , d. c. , klesman , a. j. , laird , e. s. , perez gonzalez , p. g. & mozena , m. 2011 , apj , 731 , 97 + scheuer , p. a. g. 1995 , mnras , 277 , 331 + shankar , f. , salucci , p. , granato , g. l. , de zotti , g. , & danese , l. 2004 , mnras , 354 , 1020 + silverman , j. d. , et al . 2010 , apjs , 191 , 124 + supper r. , hasinger g. , pietsch w. , trmper j. jain a. , magnier e.a . , lewin w.h.g . & van paradijs j. 1997 , a&a , 317 , 328 + swartz , d. a. , ghosh , k. k. , tennant , a. f. , & wu , k. 2004 , apjs , 154 , 519 + swartz , d. a. , soria , r. , tennant , a. f. & yukita , m. 2011 , apj , 741 , 49 + szokoly , g. p. , et al . 2004 , apjs , 155 , 271 + taylor , e. n. , et al . 2009 , apjs , 183 , 295 + timmer j. & koenig m. 1995 , a&a , 300 , 707 + tozzi , p. , et al . 2001 , apj , 562 , 42 + treister , e. , et al . 2009 , apj , 693 , 1713 + trevese , d. , boutsia , k. , vagnetti , f. , cappellaro , e. & puccetti , s. 2008 , a&a , 488 , 73 + vaughan , s. , boller , t. , fabian , a. c. , ballantyne , d. r. , brandt , w. n. , & trumper , j. 2002 , mnras , 337 , 247 + vaughan , s. , edelson , r. , warwick , r. s. & uttley , p. 2003 , mnras , 345 , 1271 + vaughan , s. , fabian , a. c. , ballantyne , d. r. , de rosa , a. , piro , l. & matt , g. 2004 , mnras , 351 , 193 + vaughan , s. & fabian , a. c. 2004 , mnras , 348 , 1415 + van den bergh , s. , herbst , e. & pritchet , c. 1973 , aj , 78 , 375 + vanzella , e. , et al . 2008 , a&a , 478 , 83 + vasudevan , r. v. , mushotzky , r. f. , winter , l. m. & fabian , a. c. 2009 , mnras , 399 , 1553 + villforth , c. , koekemoer , a. m. & grogin , n. a. 2010 , apj , 723 , 737 + walton , d. j. , roberts , t. p. , mateos , s. & heard , v. 2011 , mnras , 416 , 1844 + weisskopf , m. c. , wu , k. , trimble , v. , odell , s. l. , elsner , r. f. , zavlin , v. e. , & kouveliotou , c. 2007 , apj , 657 , 10 + xue , y. et al . 2010 , apj , 720 , 368 + xue , y. et al . 2011 , apjs , 195 , 10 + younes , g. , porquet , d. , sabra , b. & reeves , j. n. 2011 , a&a , 530 , 149 + yuan , f. & narayan , r. 2004 , apj , 612 , 724 + yun m. s. & hibbard j. e. 2001 , apj , 550 , 104 + zheng , w. , et al . 2004 , apjs , 155 , 73 + zibetti , s. , charlot , s. & rix . h. w. 2009 , mnras , 400 , 1181 + ll [ tab : columns ] 1 & sequence number in the catalog @xcite ( i.e. , xid ) + 2 & name + 3 & redshift + 4 & method of redshift measurement ( s = spectroscopic , p = photometric ) + 5 & upper limit flag for net counts + 6 & total net counts in the 0.58.0 kev band or 3@xmath24 upper limit + 7 & log @xmath42 ( ergs s@xmath13 ) + 8 - 10 & effective photon index ( @xmath25 ) and corresponding errors + 11 & variability statistic ( @xmath66 ) + 12 & probability that @xmath66 statistic is due to chance ( @xmath68 ) + 13 & maximum - to - minimum flux ratio ( @xmath78 ) + 14 - 15 & excess variance ( @xmath87 ) and corresponding error + 16 - 17 & bias - corrected excess variance ( @xmath149 , see [ sec : nxsvar ] ) and corresponding error + 18 & sfr ( @xmath105 yr@xmath13 ) + 19 & @xmath104 ( 10@xmath195 @xmath105 ) + 20 & @xmath109/@xmath110 + 21 & @xmath196 + 22 & visual galaxy classification + 23 & @xmath134 + 24 & @xmath197 + 25 & color galaxy classification + 26 & final galaxy classification + 27 & log @xmath162 ( @xmath121 ) + 28 & log @xmath165/@xmath166 + clrrrrrrrrr [ tab : galvar ] 120 & [email protected] & 1.02 & 32.7 & 41.62 & 1.4 & 2.8 & 0.0296 & 4.1 & [email protected] & [email protected] + 154 & [email protected] & 0.08 & 46.9 & 39.39 & @xmath580.9 & 2.9 & 0.0304 & 4.1 & [email protected] & [email protected] + 162 & [email protected] & 0.42 & 111.7 & 41.16 & 1.0@xmath198 & 5.8 & 0.0002 & 3.4 & [email protected] & [email protected] + 223 & [email protected] & 0.67 & @xmath2243.7 & 41.51 & @xmath581.0 & 5.4 & 0.0034 & 8.1 & [email protected] & [email protected] + 233 & [email protected] & 0.52 & 61.0 & 41.26 & @xmath581.0 & 4.4 & 0.0050 & 3.5 & [email protected] & [email protected] + ( open circles ) . galaxies with significant variability are marked as red stars . the on - axis flux limit for 20 net counts ( @xmath199 ergs @xmath12 s@xmath13 ) is plotted as a dotted line . the upper @xmath200-axis shows the maximum rest - frame timescale in years , where the maximum observed - frame timescale of the is 10.8 years . , width=528 ] ) for all sources classified as galaxies ( dark grey histogram ) and for those exhibiting significant variability ( red histogram ) . the vertical lines show the median values for all galaxies ( black , dotted ) and for variable galaxies ( red , dashed ) . a k - s test shows no significant difference between the distributions.,width=528 ] -band magnitude vs. full - band flux for sources in the main catalog of @xcite , where filled symbols mark sources with significant variability ( @xmath68 @xmath22 0.05 ) . sources classified as agn , galaxies , and stars are marked with red , black , and blue symbols , respectively . constant flux ratios are marked with diagonal lines : from top to bottom , log(@xmath201/@xmath45 ) = @xmath202 , @xmath203 , and @xmath204 . the vertical dotted line marks the on - axis full - band flux limit for 20 net counts . the sources with @xmath31-band magnitude lower limits at the top of the plot ( @xmath325% of sources ) have no @xmath31-band counterpart.,width=528 ] for sources classified as variable galaxies ( red histogram ) , with non - variable galaxies ( blue histogram ) shown for reference . as expected , the distribution of non - variable galaxies , though affected by scatter due to statistical fluctuations , centers around zero ( i.e. , variability strength is attributable to poisson fluctuations ) . the red dashed line marks the maximum variability expected from an xrb population.,width=624 ] ) vs. luminosity ( @xmath42 ) for sources classified as agn ( red circles ) and variable galaxies ( black circles and stars ) in the 4 ms main catalog , where sources with @xmath42@xmath5810@xmath127 ergs s@xmath13 are fitted with a weighted least - squares regression ( dashed line ) ; the shaded area shows the dispersion around the fitted line . black stars highlight six variable galaxies with significant xrb contribution to the total luminosity , discussed further in [ sec : suppressed ] . error bars include both measurement errors and errors associated with red - noise scatter and sampling ( [ sec : nxsvar ] ) . large blue squares mark the weighted means for each luminosity bin ; error bars are the standard deviation of the data in each bin . ultraluminous sources ( ulx ) are generally not expected above @xmath205 ergs s@xmath13 ( vertical dotted line).,width=528 ] ) vs. that expected from an xrb population ( @xmath161 ) for variable ( red stars ) and non - variable ( green squares ) galaxies . arrows designate upper limits on @xmath160 . agn ( open orange circles ) , 32 local galaxies @xcite , and 20 local luminous infrared galaxies @xcite are plotted for reference . local galaxies lie on @xmath160 = @xmath161 ( red solid line ) , with a measured dispersion of 0.34 ( red dotted lines * ? ? ? * ) .,width=528 ] ) vs. accretion rate normalized by the eddington rate ( @xmath164= @xmath165/@xmath166 ) for sources classified as agn ( red circles ) and variable galaxies ( black circles ) . a weighted least - squares regression is fitted to all variable sources ( dashed line ) , and the dotted lines show the dispersion around the best - fit line . , width=528 ] ) vs. luminosity ( @xmath42 ) as in figure [ fig : sigma_lx ] for sources with @xmath33 @xmath22 1 . model relations are shown for a universal broken power - law psd function , where the break frequency depends on @xmath162 and @xmath164 . the relations are calculated for @xmath164 = 9 @xmath6 10@xmath177 ( dashed line ) and @xmath164 = 4 @xmath6 10@xmath174 ( dash - dotted line ) , where @xmath162 = 10@xmath19110@xmath195 @xmath105 ( low to high @xmath42).,width=528 ]

the 4 ms _ chandra _ deep field - south ( ) and other deep surveys have been highly effective at selecting active galactic nuclei ( agn ) . however , cosmologically distant low - luminosity agn ( llagn ) have remained a challenge to identify due to significant contribution from the host galaxy . we identify long - term variability ( @xmath0month years , observed frame ) in 20 of 92 galaxies spanning redshifts @xmath1 that do not meet other agn selection criteria . we show that the observed variability can not be explained by binary populations or ultraluminous sources , so the variability is most likely caused by accretion onto a supermassive black hole . the variable galaxies are not heavily obscured in general , with a stacked effective power - law photon index of @xmath2 , and are therefore likely llagn . the llagn tend to lie a factor of @xmath3680 below the extrapolated linear variability - luminosity relation measured for luminous agn . this may be explained by their lower accretion rates . variability - independent black - hole mass and accretion - rate estimates for variable galaxies show that they sample a significantly different black - hole mass - accretion rate space , with masses a factor of 2.4 lower and accretion rates a factor of 22.5 lower than variable luminous agn at the same redshift . we find that an empirical model based on a universal broken power - law psd function , where the break frequency depends on smbh mass and accretion rate , roughly reproduces the shape , but not the normalization , of the variability - luminosity trends measured for variable galaxies and more luminous agn .

let @xmath0 be an odd prime , and let @xmath2 be the finite cyclic group @xmath1 . we assume that all 3-manifolds are compact and closed . the quantum invariant of a 3-manifold can be defined using any modular category . in @xcite , gilmer was interested in studying the relation between the @xmath3- invariants of @xmath0-periodic 3-manifolds and their quotient manifolds . he obtained a congruence relating these invariants . his result was obtained by using the trace formula of topological quantum field theory ( see proposition ( [ p : t ] ) ) and studying gaussian sums . chbili used the results about the jones polynomial and the kauffman multi - bracket of @xmath0-periodic links to obtain a similar result for rational homology 3-spheres for the @xmath4-invariants in @xcite . also in @xcite , he gave similar results for the @xmath5 and the moo - invariants . moreover in @xcite , chen and le generalized the above results for rational homology spheres using any complex simple lie algebra . we give similar results for all 3-manifolds using any modular category over an integrally closed ground ring . our proof takes place completely in the context of modular categories . we use the surgery descriptions of @xmath0-periodic 3-manifold and its orbit manifold , obtained in @xcite , to prove the result . in section [ 1 ] , we give a brief exposition on how to calculate the quantum invariant for any 3-manifold from its surgery description . in section [ 2 ] , we discuss the @xmath1-actions on 3-manifolds and the relation between the link that describes a @xmath0-periodic 3-manifold and the link that describes its orbit manifold . some formulas and results regarding the value of colored ribbon graphs under the covariant functor @xmath6 will be given in section [ 3 ] . finally in section [ 4 ] , we state and prove the main result . fix a strict modular category ( @xmath7,@xmath8 ) with ground ring @xmath9 and a rank @xmath10 . a result due to lickorish and wallace asserts that every closed oriented 3-manifold can be obtained by surgery on @xmath11 along a framed link . let @xmath13 be a closed oriented 3-manifold obtained by surgery on @xmath11 along a framed link @xmath14 . the @xmath12-invariant of @xmath15 associated to ( @xmath16 ) where @xmath17 is a colored ribbon graph in @xmath13 is given by @xmath18 here @xmath19 is the signature of the linking matrix of the link @xmath14 , and @xmath20 is the number of components of @xmath14 , and @xmath21 where @xmath22 denotes the diagram for the unknot with a single double point and writhe -1 . we use the notation @xmath23 where col(@xmath14 ) is the set of all mappings from the set of components of @xmath14 to @xmath24 ( the set of simple objects ) , and @xmath25 is the ribbon graph obtained by coloring the @xmath26-th component of @xmath14 by @xmath27 . here @xmath6 is the covariant functor defined in ( * ? ? ? * chapter.i ) which assigns to a @xmath7-colored ribbon graph in @xmath28 an element of the ground ring . the material of this section is due to turaev @xcite . we take the definition of the quantum invariant to be as follows @xmath29 our result is simpler when expressed using this normalization . [ t : t][turaev ] @xmath30 ( or @xmath31 ) is a topological invariant of the pair @xmath15 . * example . * we know that @xmath11 is obtained by doing surgery on the empty link , i.e @xmath32 . also , @xmath11 is obtained by doing surgery along the hopf link @xmath33 with framing 0 on both components . hence , we conclude @xmath34 . @xmath35 is invariant under kirby sliding . this corollary is really major part of proof of theorem ( [ t : t ] ) . finally , the @xmath12-invariant can be recovered in terms of the tqft - theory @xmath36 which is a functor from the category @xmath37 whose objects are closed surfaces and 3-manifolds as its morphisms ( the surfaces and 3-manifolds have banded links sitting inside of them ) to the category of @xmath9-modules and @xmath9-linear homomorphisms , where @xmath38 is the functor on the surfaces and @xmath39 is the functor on 3-manifolds . in fact , the assigned value of a closed 3-manifold under @xmath39 is a scalar multiplication homomorphism from the base ring to itself and that scalar is the @xmath12-invariant of that manifold . for more details of tqft see ( @xcite ) let @xmath13 be a closed oriented 3-manifold that is a result of surgery on @xmath11 along the framed link @xmath14 . a framed link @xmath14 in @xmath11 is said to be @xmath0-periodic if there exists a @xmath1-action on @xmath11 , with a fixed point set equal to a circle , that maps @xmath14 to itself under this action and @xmath14 is assumed to be disjoint from the circle . @xmath13 is said to be @xmath0-periodic if there is an orientation preserving @xmath1-action with fixed point set equal to a circle , and the action is free outside this circle . now we list the following two results from @xcite that will be used in later sections . [ t : nonfree ] there is a @xmath1-action on @xmath13 with a fixed point set equal to a circle iff @xmath13 can be obtained be as a result of surgery on a @xmath0-periodic link @xmath14 and @xmath1 acts freely on the set of the components of @xmath14 . by the positive solution of the smith conjecture , we can represent any framed @xmath0-periodic link as a closure of some graph such that the rotation of this graph about the @xmath40-axis in @xmath41 ( or the circle in @xmath11 ) by 2@xmath42 leaves it invariant , i.e @xmath43 ( where the bar means the closure of the graph ) see figure [ f : figure1 ] . + let @xmath44 denote the orbit space , then @xmath45 is obtained by surgery on @xmath11 along the link @xmath46 . [ l : basic ] let @xmath14 a @xmath0-periodic link in @xmath11 . the following are equivalent 1 . @xmath1 acts freely on the set of components of @xmath14 ; 2 . the linking number of each component of the @xmath47 the axis of the action is congruent to zero modulo @xmath0 ; 3 . the number of components of @xmath14 is equal to @xmath0 times the number of components of @xmath47 . we use two different notions of trace one is the trace of a linear homomorphism ( denoted by @xmath48 ) in the category of @xmath9-modules and the other one is the trace of a ribbon graph ( denoted by @xmath49 ) in the category of ribbon graphs defined in ( * ? ? ? * chapter.1 ) [ p : t ] @xmath50 where @xmath17 is a colored ribbon @xmath51 tangle in @xmath52 . this is a special case of the trace formula for tqft ( * ? ? ? * prop.1.2 ) and ( * ? ? ? * ex.2.8.1 ) . [ l : newtrace ] @xmath53 let @xmath33 stands for the zero - framed hopf link on both components . we have @xmath54{t94}}\\ & = \frac{1}{\{h\}}\sum_{\lambda \in col(h ) } \{h\}_{\lambda } f(\overline{\omega } ) \ , \quad \text{as}\ \{h\ } = \sum_{\lambda\in col(l ) } \{h\}\\ & = \frac{1}{\mathcal{d}^{2}}\sum_{\lambda \in col(h ) } \dim(\lambda)f(\overline{\omega}\cup h)\ , \quad\text{as}\ \{h\ } = \mathcal{d}^{2},\ \text{where } \ h \ \text{is unlinked from}\ \overline{\omega } \\ & = \frac{1}{\mathcal{d}^{2}}\sum_{\lambda \in col(h ) } \dim(\lambda)f(\overline{\omega^ { ' } } ) \ , \quad \substack{\text{using the invariance of } \ \{l,\omega\ } \ \\ \text{under sliding see figure}\ref{f : figure2 } } \\ & = \sum_{i \in i } \dim(v_{i})\mathcal{d}^{-2}\sum_{j \in i}\dim(v_{j } ) f ( \overline{\omega^{'}})\\ & = \sum_{i \in i } \dim(v_{i } ) \tau(s^{2}\times s^{1},\overline{1_{v_{i}}\otimes \omega})\ , \quad \text{by formula ( \ref{e : rtc } ) } \\ & = \sum_{i \in i } \dim(v_{i } ) \operatorname{trace}_{v(s^{2},l+1)}(z_{(s^{2}\times i,{1_{v_{i}}\otimes \omega})})\ , \quad \text{by proposition ( \ref{p : t } ) } \end{split}\ ] ] obtained by sliding @xmath55 over the hopf link using the second kirby move , title="fig:",width=288 ] + let @xmath56 be the ideal generated by @xmath0 and @xmath57 in @xmath9 . [ c : trace ] let @xmath17 be any colored ribbon graph over any modular category with integrally closed ground ring . then @xmath58 if we assume @xmath59 it follows that @xmath60 hence the result follows from lemma ( [ l : newtrace ] ) and ( * ? ? ? * lem.3.5(i ) ) , which implies that @xmath61^{p } \quad \pmod { p},\ ] ] where @xmath39 is an endomorphism of free @xmath9-module . let @xmath13 be a 3-manifold that admits a @xmath1-action with a fixed point set equal to a circle . then we are in situation of theorem [ t : nonfree ] . i.e. @xmath13 is obtained by surgery on @xmath11 along a framed @xmath0-periodic link @xmath14 ( see figure [ f : figure1 ] ) . we would like to relate the quantum invariant of @xmath13 to the quantum invariant of @xmath44 . before we do so , we introduce the following . let @xmath14 be a @xmath0-periodic link , and @xmath62 be a coloring of @xmath14 . if @xmath25 is invariant under the rotation of the graph that represents @xmath14 by 2@xmath42 , then @xmath62 is called a @xmath0-periodic coloring . let us start with any coloring of @xmath14 say @xmath62 , either @xmath62 is @xmath0-periodic or not . let us assume that @xmath62 is not @xmath0-periodic , i.e @xmath25 is not invariant under the rotation by 2@xmath42 about the @xmath40-axis . hence the @xmath26-th rotation of @xmath25 ( the rotation by @xmath64 ) represents a ribbon graph with the same value under @xmath6 ( since @xmath6 is an isotopy invariant ) and different coloring denoted by @xmath65 . so the term with a non - periodic coloring occurs @xmath0 times . hence we reduce the summation on the left - hand side to the periodic colorings . now the result follows from corollary ( [ c : trace ] ) and the fact that the periodic colorings of @xmath14 are in one - to - one correspondence with the colorings of @xmath47(by restriction ) . suppose @xmath9 is a knot in a homology sphere @xmath45 . let @xmath73 be the @xmath0-fold cyclic cover branched along @xmath9 . it is known that , we can extend this cover to a cover @xmath74 of 4-manifolds ( where @xmath75 and @xmath76 ) branching over the surface @xmath77 . let @xmath78 denote the self - intersection of @xmath77 in @xmath79 using the framing on @xmath9 obtained from any seifert surface for @xmath9 in @xmath45 . in this case , we define the total signature @xmath80 @xmath81 by a well - known argument , using novikov additivity and the @xmath2-signature theory @xcite , @xmath80 is independent of the choices made . the linking matrix of @xmath14 describes the intersection form on 4-manifold with boundary @xmath13 which is a branched cover along a disk with zero self - intersection in a 4-manifold with boundary @xmath45 . the corollary now follows by identifying @xmath68 with the total signature of @xmath9 . the results of this paper were obtained during my ph.d . studies at louisiana state university . i would like to express deep gratitude to my advisor dr.patrick gilmer whose comments and support were very important for the completion of this project . w c. blanchet , n. habegger , g. masbaum , and p. vogel , _ topological quantum field theories derived from the kauffman bracket _ , topology * 34 * ( 1995 ) , 883 - 927 . n. chbili , _ quantum invariants of finite group actions on three manifolds _ , topology and its applications * 136 * ( 2004 ) , 219 - 231 . n. chbili , _ les invariants @xmath84 des 3-vari@xmath85t@xmath85s p@xmath85riodiques _ , ann . fourier * 51*(2001 ) 1135 - 1150 . q. chen , t. le , _ quantum invariants of periodic links and periodic 3-manifolds _ , fund . * 184 * ( 2004 ) , 5571 . m. gilmer , _ quantum invariants of periodic three - manifolds _ , geometry and topology monographs * 2 * ( 1999 ) , 157175 . m. gilmer , j. kania - bartoszynska , and j. h. przytycki , _ 3-manifold invariants and periodicity of homology spheres _ , algebraic and geometric topology * 2 * ( 2002 ) , 825842 . l. kauffman , _ on knots _ , number 115 , annals of mathematics studies , princeton university press , 1987 . j. h. przytycki and m. v. sokolov , _ surgeries on periodic links and homology of periodic 3-manifolds _ , math . soc * 131 * ( 2001 ) , 295307 .

a @xmath0-periodic 3-manifold is a 3-manifold that admits a @xmath1-action whose fixed point set is a circle . we give a congruence that relates the quantum invariant of a @xmath0-periodic 3-manifold associated to any modular category over an integrally closed ground ring and the corresponding quantum invariant of its orbit space .

the permeability of porous media is an important parameter to predict the unconventional gas production . for laminar flows in highly permeable porous media , the darcy s law states that the volume flow rate @xmath0 is proportional to the pressure gradient @xmath1 : @xmath2 where @xmath3 is the cross - section area of the flow , @xmath4 is the shear viscosity of the fluid , and @xmath5 is the permeability of a porous medium that is independent of the fluid . for this reason , @xmath5 is known as the intrinsic permeability . for gas flows in low permeable porous media , however , the measured permeability is larger than the intrinsic permeability and increases with the reciprocal mean gas pressure @xmath6 @xcite . in order to distinguish it from the intrinsic permeability , the permeability is called the apparent permeability @xmath7 , which can be expressed as : @xmath8 where @xmath9 is the correction factor . the variation of the apparent permeability with respect to the mean gas pressure is due to the rarefaction effects , where infrequent collisions between gas molecules not only cause the gas slippage at the solid surface , but also modify the constitution relation between the stress and strain - rate @xcite . the extent of rarefaction is characterized by the knudsen number @xmath10 ( i.e. the ratio of the mean free path @xmath11 of gas molecules to the characteristic flow length @xmath12 ) : @xmath13 where @xmath14 is the shear viscosity of the gas at a reference temperature @xmath15 , and @xmath16 is the gas constant . gas flows can be classified into four regimes ; for gas flows in porous media , the region of @xmath10 for different flow regimes may change . ] : continuum flow @xmath17 in which navier - stokes equations ( nses ) can be used ; slip flow @xmath18 where nses with appropriate velocity - slip / temperature - jump boundary conditions may be used ; transition flow @xmath19 and free - molecular flow @xmath20 , where nses break down and the boltzmann equation is used to describe rarefied gas flows @xcite . recently , based on nses with the first - order velocity - slip boundary condition ( fvbc ) , @xcite found that the apparent gas permeability of the porous media is a nonlinear function of @xmath10 . this result , however , is questionable , because the nses were used beyond its validity . through our theoretical analysis and numerical calculations , we show that nses with fvbc can only predict the apparent permeability of porous media to the first - order accuracy of @xmath10 . consider a gas flowing through the periodic porous media . suppose the geometry along the @xmath21-direction is uniform and infinite , the gas flow is effectively two - dimensional and can be studied in a unit rectangular cell abcd , with appropriate governing equations and boundary conditions ; one example of the porous medium consisting of a periodic array of discs is shown in fig . we are interested in how the apparent permeability varies with the knudsen number . the boltzmann equation is fundamental in the study of rarefied gas dynamics from the continuum to the free - molecular flow regimes , which uses the distribution function @xmath22 to describe the system state : @xmath23 where @xmath24 is the three - dimensional molecular velocity normalized by the most probable speed @xmath25 , @xmath26 is the spatial coordinate normalized by the length @xmath12 of the side ab , @xmath27 is the time normalized by @xmath28 , @xmath29 is normalized by @xmath30 , while @xmath31 is the boltzmann collision operator . in order to save the computational cost , @xmath31 is usually replaced by the relaxation - time approximation @xcite , resulting in the bhatnagar - gross - krook ( bgk ) equation . numerical simulation for the poiseuille flow between two parallel plates shows that the bgk equation can yield accurate mass flow rates when the gas flow is not in the free - molecular regime @xcite . when the porous medium is so long that the pressure gradient is small , the bgk equation can be linearized . the distribution function is expressed as @xmath32 , where @xmath33 is the equilibrium distribution function , and the perturbation @xmath34 is governed by @xcite : @xmath35,\ ] ] where macroscopic quantities such as the perturbed density @xmath36 , the velocity @xmath37 and @xmath38 , and the perturbed temperature @xmath39 are calculated as @xmath40 the kinetic equation has to be supplied with the boundary condition . suppose the pressure gradient is along the @xmath41 direction , on the inlet and outlet of the computational domain abcd ( the coordinates of the four corners a , b , c , and d are @xmath42 , and @xmath43 , respectively ) , the pressure gradient is applied and the periodic condition for the flow velocity is used @xcite : @xmath44 at the lines ab and cd , the specular reflection boundary condition is used to account for the symmetry : @xmath45 when @xmath46 , and @xmath47 when @xmath48 , while at the solid surface , the diffuse boundary condition is used @xcite : @xmath49 where @xmath50 is the normal velocity vector at the solid surface . the apparent gas permeability , which is normalized by @xmath51 , is calculated by @xmath52 historically , the state of a gas is first described by macroscopic quantities such as the density @xmath53 , velocity @xmath54 , and temperature @xmath55 ; and its dynamics is described by the euler equations or nses ( based on the empirical newton s law for stress and the fourier s law for heat flux ) . these equations , however , can be derived rigorously from the boltzmann equation , at various order of approximations . by taking the velocity moments of the boltzmann equation , the five macroscopic quantities are governed by the following equations : @xmath56 however , the above equations are not closed , since expressions for the shear stress @xmath57 and heat flux @xmath58 are not known . one way to close - is to use the chapman - enskog expansion , where the distribution function is expressed in the power series of @xmath10 @xcite : @xmath59 where @xmath60 is the equilibrium maxwellian distribution function . when @xmath61 , we have @xmath62 , and - reduce to the euler equations . when the distribution function is truncated at the first - order of @xmath10 , that is , @xmath63 we have @xmath64 and - reduce to nses . when @xmath65 , burnett equations can be derived . alternatively , following the method of @xcite , 13- , 20- and 26-moment equations @xcite can be derived from the boltzmann equation to describe flows at different levels of rarefaction . here the regularized 20-moment ( r20 ) equations are used , which , in addition to - , include governing equations for the high - order moments @xmath57 , @xmath58 , and @xmath66 : @xmath67 @xmath68 @xmath69 the constitutive relationships between the unknown higher - order moments ( @xmath70 , @xmath71 and @xmath72 ) and the lower - order moments were given by structrup & torrilhon ( 2003 ) and gu & emerson ( 2009 ) to close to . for linearized flows , it is adequate to use the linear gradient transport terms only and they are : @xmath73 where the collision constant @xmath74 is 2.097 for maxwell molecules @xcite . macroscopic wall boundary conditions were obtained from the diffuse boundary condition @xcite . in a frame where the coordinates are attached to the wall , with @xmath75 the normal vector of the wall pointing towards the gas and @xmath76 the tangential vector of the wall , the velocity - slip parallel to the wall @xmath77 and temperature - jump conditions are : @xmath78 where @xmath79 . the rest of wall boundary conditions for higher - order moments are listed in appendix [ wall_hob ] . note that the velocity - slip boundary condition is also of higher - order due to the appearance of the higher - order moment @xmath80 . following the above introduction , it is clear that nses with fvbc are only accurate to the first - order of @xmath10 ; therefore , any apparent gas permeability showing the nonlinear dependence with @xmath10 is highly questionable . the r20 equations are accurate to the third - order of @xmath10 @xcite , which should give the some apparent permeability of the porous media as nses when @xmath81 , and be more accurate than nses as @xmath10 increases . numerical simulations are also performed to demonstrate this . we first investigate the rarefied gas through a periodic array of discs with the diameter @xmath82 , as shown in fig . [ geo ] . using nses and fvbc , when the porosity @xmath83 is large , the slip - corrected permeability can be obtained analytically @xcite : @xmath84,\ ] ] where @xmath85 is the solid fraction and @xmath86 when the diffuse boundary condition is used @xcite . the intrinsic permeability is obtained when @xmath87 . + the accuracy of the slip - corrected permeability is assessed by comparing to numerical solutions of the bgk equation and r20 equations , when the porosity is @xmath88 . for the linearized bgk equation , two reduced distribution functions were introduced to cast the three - dimensional molecular velocity space into a two - dimensional one , and the obtained two equations are solved numerically by the discrete velocity method @xcite and the unified gas kinetic scheme @xcite . in the unified gas kinetic scheme , a body - fitted structured curvilinear mesh is used , with 150 lines along the radial direction and 300 lines along the circumferential direction , see fig . [ apparent_permeability](a ) . in the discrete velocity method , a cartesian grid with @xmath89 equally - spaced points is used and the solid surface is approximated by thestair - case " . in solving the r20 equations , a similar body - fitted mesh with @xmath90 cells is used , and the detailed numerical method is given by @xcite . the molecular velocity space in the bgk equation is also discretized : @xmath91 and @xmath92 are approximated by the @xmath93 gauss - hermite quadrature when @xmath10 is small ( @xmath94 in this case ) , and the newton - cotes quadrature with @xmath95 non - uniform discrete velocity points @xcite when @xmath10 is large . the apparent permeability is plotted in fig . [ apparent_permeability](b ) as a function of @xmath10 . when @xmath96 , our numerical simulations based on the linearized bgk equation and r20 equations agree with each other , and the apparent permeability is a linear function of @xmath10 . when @xmath97 , the r20 equations , although being accurate to the third - order of @xmath10 , predict lower apparent permeability than that of the bgk equation . the slip - corrected permeability increases linearly with @xmath10 only when @xmath98 , and then quickly reaches to a maximum value when @xmath97 . this comparison clearly demonstrates that , nses with fvbc are only accurate to the first - order of @xmath10 . this result is in accordance with the approximation adopted in the derivation of nses from the boltzmann equation . although the `` curvature of the solid - gas interface '' makes the apparent permeability a concave function of @xmath10 in the framework of nses with fvbc @xcite , higher - order moments in - and the higher - order velocity slip in , which are derived from the boltzmann equation and the gas kinetic boundary condition to the third - order accuracy of @xmath10 , restore linear dependence of the apparent permeability on @xmath10 when @xmath96 . + the conclusion that nses with fvbc is accurate only to the first - order of @xmath10 not only holds for the simple porous medium as shown in fig . [ geo ] , but also applies to more complex porous media , for example , see the unit cell in fig . [ randomdisc](a ) where the porosity is 0.6 . in this case , the linearized bgk equation is solved by the discrete velocity method , with a cartesian mesh of @xmath99 cells ; the grid convergence is verified , as using @xmath100 cells only results in a 0.6% increase of the apparent permeability when @xmath101 . nses with fvbc are solved in openfoam using the simple algorithm and a cell - centered finite - volume discretization scheme , on unstructured grids . a body - fitted computational grid is generated using the openfoam meshing tool , resulting in a mesh of about 600,000 cells of which the majority are hexahedra and the rest few close to the walls are prisms . the apparent permeability from nses increases linearly over a very narrow region of the knudsen number ( i.e. @xmath102 ) and then quickly reaches a constant value . again , from fig . [ randomdisc](b ) we see that nses with fvbc is roughly accurate when the apparent permeability is a linear function of @xmath10 ; in this region of @xmath10 , the maximum apparent permeability @xmath7 is only about one and a half times larger than the intrinsic permeability @xmath5 . in summary , through our numerical simulations based on the linearized bhatnagar - gross - krook equation and the regularized 20-moment equations , we show that the navier - stokes equations with the first - order velocity - slip boundary condition can only predict the apparent permeability of the porous media to the first - order accuracy of the knudsen number . lw acknowledges the support of an early career researcher international exchange award from the glasgow research partnership in engineering , allowing him to visit the hong kong university of science and technology for one month . lg thanks lianhua zhu for helpful discussions on the first - order velocity - slip boundary condition in openfoam . this work is also partly supported by the engineering and physical sciences research council in the uk under grant ep / m021475/1 . the wall boundary conditions for higher - order moments are given as follows : @xmath103 where @xmath104 and @xmath105 .

in a recent paper by lasseux , valds - parada and porter ( j. fluid mech . * 805 * ( 2016 ) 118 - 146 ) , it is found that the apparent gas permeability of the porous media is a nonlinear function of the knudsen number . however , this result is highly questionable , because the adopted navier - stokes equations and the first - order velocity - slip boundary condition are first - order ( in terms of the knudsen number ) approximations of the boltzmann equation and the kinetic boundary condition for rarefied gas flows . our numerical simulations based on the bhatnagar - gross - krook kinetic equation and regularized 20-moment equations prove that the navier - stokes equations with the first - order velocity - slip boundary condition are only accurate at a very small knudsen number limit , where the apparent gas permeability is a linear function of the knudsen number . authors should not insert the keywords

in recent times , lot of scientific interests have been placed on ir - based 5@xmath0 transition metal oxides ( tmos ) . these materials have enhanced spin - orbit coupling ( soc ) effect and sizable crystal field effect ( cfe ) due to presence of heavy ir atoms . on the other hand , the electronic correlation effect ( @xmath4 ) , which has reasonable effect in 3@xmath0 based tmos , is much weak in these materials due to extended character of 5@xmath0 orbitals . the complex interplay between these competing interactions such as , soc , @xmath4 and cfe is believed to give rise @xmath5 = 1/2 electronic state which in turn stabilizes various exotic ground states in these ir - based oxide materials.@xcite the structural organization in iridium oxides plays crucial role for their magnetic and electronic states . following general ruddlesden - popper series sr@xmath6ir@xmath7o@xmath8 , the @xmath9 = 1 realizes layered compound sr@xmath1iro@xmath2 which is highly insulating and has canted - type antiferromagnetic ( afm ) state at low temperature.@xcite the @xmath9 = 2 gives bi - layered compound sr@xmath10ir@xmath1o@xmath11 which also has an insulating state and antiferromagnetic type spin ordering at low temperature , however , resistivity shows much moderate value compared to its @xmath9 = 1 counterpart.@xcite in contrast , the sriro@xmath10 which is stabilized for @xmath9 = @xmath12 , shows a metallic ground state with paramagnetic behavior.@xcite nonetheless , insulating phases in iridium oxides appear to be related with the nature of magnetic state of the materials . recently , an intriguing coupling between structure , magnetism and electrical conductivity has been shown for sr@xmath1iro@xmath2.@xcite therefore , understanding the nature of magnetism is much necessary to comprehend the nature of insulating phase in these materials . the k@xmath1nif@xmath2-based layered compound sr@xmath1iro@xmath2 has drawn special interest as recent theoretical calculations have predicted for possible superconductivity in this material due to its structural similarity with superconducting la@xmath1cuo@xmath2 and sr@xmath1ruo@xmath2 materials.@xcite on cooling , sr@xmath1iro@xmath2 exhibits paramagnetic ( pm ) to canted antiferromagnetic ( afm ) transition around 230 k. the successive rotation of iro@xmath13 octahedra ( @xmath1411.3@xmath15 along c - axis ) induces dzyaloshinsky - moriya ( dm ) type antisymmetric exchange interaction which in turn gives a weak ferromagnetic ( fm ) behavior below the magnetic ordering temperature . the insulating state in sr@xmath1iro@xmath2 is rather much debated , and has been largely explained with the description of mott s and slater s theory . following the former picture , the 5@xmath16 electronic state of ir@xmath17 completely fills the @xmath5 = 3/2 and partially fills the @xmath5 = 1/2 state . the @xmath5 = 1/2 has very narrow bandwidth which in presence of even small @xmath4 splits up , and opens a mott - like gap.@xcite in this picture , insulating phase is driven by an electronic correlation effect being independent of appearance of magnetic ordering though this model favors the localized picture of spin interaction . in the picture of slater model , on other hand , insulating phase is a consequence of magnetic transition as the long - range type antiferromagnetic ordering results in opening of band gap at the boundaries of folded brillouin zone.@xcite this branch of insulator basically has mean - field type magnetic interaction with itinerant character . both theoretical as well as experimental evidences have been put forward in support of both these insulating phases in sr@xmath1iro@xmath2 . the @xmath5 = 1/2 dominated mott insulating phase has been discussed with the theoretical models@xcite and also been verified from experiments such as , angel resolved photoemission spectroscopy ( arpes)@xcite and resonant x - ray scattering ( rxs)@xcite experiments . in parallel developments , a recent study by x - ray absorption spectroscopy have shown a deviation from strong soc dominated @xmath5 = 1/2 state in sr@xmath1iro@xmath2.@xcite additionally , a recent neutron diffraction measurement has shown moment of ir in sr@xmath1iro@xmath2 to be 0.208 @xmath18 which appears much low compared to the calculated spin - only value 1 @xmath18/ir - site for localized spin of @xmath19 = 1/2.@xcite while there is no clear experimental evidences for metallic behavior in paramagnetic state of sr@xmath1iro@xmath2 , the band structure calculations by arita _ _ et al.__@xcite have clearly predicted sr@xmath1iro@xmath2 to be a slater - type insulator . another calculation by watanabe _ _ et al.__@xcite has shown sr@xmath1iro@xmath2 has moderate electronic correlation effect and basically lies in between slater- and mott - type insulating phase . on experimental side , a time - resolved optical study has shown an unique coexistence of slater- and mott - type insulating phase in sr@xmath1iro@xmath2.@xcite a recent study with scanning tunneling microscopy has interestingly shown that insulating gap in sr@xmath1iro@xmath2 opens up around the magnetic transition and the nature of insulating phase in this material is mostly of slater - type.@xcite following this controversy , we have attempted to understand the nature of insulating phase in sr@xmath1iro@xmath2 by means of understanding the nature of magnetic interaction . in that connection , we have estimated the critical exponents , analyzed the thermal demagnetization data and studied the magnetocaloric effect ( mce ) . the estimated exponents while do not exactly match with the universality classes but values are close to the mean - field interaction model . analysis of thermal demagnetization data at low temperature , however , imply dual presence of localized and itinerant model of magnetization . the mce experiment suggest magnetic interaction in sr@xmath1iro@xmath2 is of mean - field type . from these , we infer that insulating phase in sr@xmath1iro@xmath2 has contribution from both slater and mott mechanism but it is more close to the slater one . the single - phase polycrystalline sample has been prepared using solid state reaction method . details of sample preparation and characterization are given in elsewhere.@xcite the magnetization ( @xmath20 ) data as a function of temperature ( @xmath21 ) and magnetic field ( @xmath22 ) have been collected using physical property measurement system ( ppms ) by quantum design . for critical analysis , magnetic isotherms are collected across @xmath3 in temperature range from 218 to 238 k with a temperature step @xmath23 = 2 k. to calculate the change in magnetic entropy @xmath24 , the @xmath25 data have been collected in temperature range 5 to 300 k up to field 70 koe . temperature dependent magnetization data collected following zero field cooled ( zfc ) and field cooled ( fc ) protocol in applied field of 10 koe are shown in fig . 1 . with decreasing temperature , both branches of magnetization shows steep rise below 250 k which is marked by pm to fm phase transition . at low temperature below about 95 k , the @xmath26 shows a sharp fall which has been understood to arise due to complex magneto - structural coupling in this material.@xcite this material otherwise makes transition from pm to afm state around 230 k. the weak fm behavior is realized a result of spin canting in afm state which arises due to dzyaloshinsky - moriya ( dm ) type interaction induced by rotation of iro@xmath13 octahedra . in deed , a trace of spontaneous magnetization at low temperature has earlier been observed in sr@xmath1iro@xmath2.@xcite vs @xmath22 ) measured at different temperatures are shown across @xmath3 for sr@xmath1iro@xmath2.,width=302 ] the sr@xmath1iro@xmath2 has magnetic transition from pm to fm state which is of second order . across the second order phase transition the spontaneous magnetization ( @xmath27 ) below @xmath3 , inverse magnetic susceptibility ( @xmath28 ) above @xmath3 and the magnetization at @xmath3 follow power - law behavior with temperature as given below : @xmath29 @xmath30 @xmath31 vs @xmath32 ) as obtained from isotherms are shown at different temperatures for sr@xmath1iro@xmath2.,width=302 ] where @xmath33 = ( @xmath21 - @xmath3)/@xmath3 is the reduced temperature . the @xmath34 , @xmath35 and @xmath36 are the critical amplitudes and @xmath37 , @xmath38 and @xmath39 are the critical exponents . magnetic phase transitions are generally classified into different universality classes based on the dimensionality of lattice system ( @xmath0 ) and the dimensionality of spin system ( @xmath9 ) . in principle , critical exponents in asymptotic regime ( @xmath33 @xmath40 0 ) exhibit universal values depending on the universality class the system belong to , where the exponents are independent of microscopic details of the system . to understand the nature of magnetic state in sr@xmath1iro@xmath2 we have estimated the exponents by measuring magnetic isotherms @xmath25 around the @xmath3 . the exponents have been estimated following different methods such as , arrott plot,@xcite critical isotherm analysis and kouvel - fisher analysis.@xcite fig . 2 shows @xmath25 plots measured in temperature range from 218 to 238 k at an interval of 2 k. the general shape of @xmath20 vs @xmath22 in main panel and the decreasing slope of @xmath41 vs @xmath22 in inset of fig . 2 implies a second - order type phase transition in sr@xmath1iro@xmath2 . to further analyze the magnetization data we have plotted @xmath25 isotherms in terms of arrott plot@xcite i.e. , @xmath42 vs @xmath32 . in general , the system which follows mean - field spin interaction model ( @xmath37 = 0.5 and @xmath38 = 1.0 ) the arrott plot forms a set of parallel straight lines . the slope taken in higher field regime in arrott plot directly gives @xmath27 and @xmath43 as an intercept on positive @xmath42 and @xmath32 axis , respectively . moreover , the isotherm which passes through origin in arrott plot gives @xmath3 of the material . 3 shows such arrott plot of material sr@xmath1iro@xmath2 with the @xmath25 data from fig . 2 . as clear in fig . 3 , isotherms are not perfectly parallel straight lines and a minor tuning of exponents are required for parallel straight lines . nonetheless , fig . 3 suggests the magnetic interaction in sr@xmath1iro@xmath2 is close to mean - field type . it is worth mentioning that we have tried other relevant models for localized spin interaction such as , 3-dimensional heisenberg and 3-dimensional ising model but the resulting isotherms do not form set of parallel straight lines . to find out right values of critical exponents we have used generalized modified arrott plot:@xcite @xmath44 = 0.55 and @xmath38 = 1.15 for sr@xmath1iro@xmath2.,width=302 ] where the @xmath45 and @xmath46 are the constants . in eq . 4 , the mean - field exponents @xmath37 = 0.5 and @xmath38 = 1.0 recover the original arrott plot . as the eq . 4 contains two unknown parameters @xmath37 and @xmath38 , therefore tunning these parameters to get parallel straight lines is a nontrivial task . however , we have followed the rigorous iterative process where the @xmath27 and @xmath47 obtained from fig . 3 are used in eqs . 1 and 2 as an input , and obtained new @xmath37 and @xmath38 values have been used to construct figure similar to fig . this process has been continued till stable values of @xmath37 and @xmath38 are obtained . 4 shows modified arrott plot with the so obtained exponents @xmath37 = 0.55 and @xmath38 = 1.15 . it is evident in figure that a set of parallel straight lines is obtained in higher field regime . the lines in lower field regime , however , are curved as the measured magnetization is an average value of contributions from different domains which are magnetized in different directions . we find that isotherm taken at temperature 225 k passes through origin which is the @xmath3 of this material . the obtained @xmath3 ( @xmath14 225 k ) in present study conforms with the value determined from neutron diffraction study@xcite ( 224(2 ) k ) and resonant magnetic x - ray diffuse scattering study@xcite ( 228.5(0.5 ) k ) . for further checking , the @xmath27 and @xmath47 obtained from intercepts on @xmath48 vs @xmath49 axises in fig . 4 , have been plotted in fig . it is evident in figure that @xmath50 and @xmath51 reasonably follows eqs . 1 and 2 . the solid lines are due to fitting with eqs . 1 and 2 , respectively . the fitting yields @xmath37 = 0.557(1 ) and @xmath3 = 225.1 k and @xmath38 = 1.129(1 ) and @xmath3 = 224.9 k. the obtained values are pretty close to the values in fig . 4 . ( left axis ) and inverse initial susceptibility @xmath47 ( right axis ) determined from the linear fitting of modified arrott plot in fig . 4 are shown for sr@xmath1iro@xmath2 . lines are due to fitting with eqs . 1 and 2.,width=302 ] ( left axis ) and inverse initial susceptibility @xmath47 ( right axis ) are shown for sr@xmath1iro@xmath2 . solid lines are due to linear fitting of data.,width=302 ] the critical exponents as well as temperature have also been obtained using kouvel - fisher ( kf ) method.@xcite this method shows following relationships : @xmath52 @xmath53 [ cols="^,^,^,^,^,^",options="header " , ] the eqs . 5 and 6 imply that slopes obtained from plotting of @xmath54 vs @xmath21 and @xmath55 vs @xmath21 will give 1/@xmath37 and 1/@xmath38 , respectively . moreover , @xmath3 can be correctly and independently obtained from the intersection point on temperature axis . 6 shows kouvel - fisher plot with the @xmath27 and @xmath47 values obtained from fig . 6 shows a reasonable straight line fitting which yields @xmath37 = 0.551(1 ) and @xmath3 = 225.3 k and @xmath38 = 1.143(1 ) and @xmath3 = 225.1 k. the values obtained following kouvel - fisher plot are quite consistent with the ones obtained from modified arrott plot in fig . 4 . to determine the exponent @xmath39 we have used eq . 3 which shows plotting of @xmath25 at @xmath3 in the form of @xmath56 vs @xmath57 would produce straight line with slope 1/@xmath39 . 7 shows critical isotherm @xmath25 at @xmath3 = 225 k. inset shows same plot in log - log scale where a straight line fitting yields @xmath39 = 3.08(2 ) . the obtained @xmath39 is close to the value ( 3.0 ) predicted for mean - field interaction model . the exponent @xmath39 has also been obtained using widom scaling equation which predicts a relationship among the exponents @xmath37 , @xmath38 and @xmath39 as following,@xcite @xmath58 using eq . 7 , the exponent @xmath39 has been calculated with the @xmath37 and @xmath38 obtained from modified arrott plot and kouvel - fisher plot . table i shows the calculated values of @xmath39 agree well with the value obtained in fig . 7 , thus confirming the authenticity of the determined exponents . inset shows log - log plotting of same data and the solid line is due to linear fitting.,width=264 ] the critical exponents ( @xmath37 , @xmath38 and @xmath39 ) and the critical temperature ( @xmath3 ) have been obtained from different self - consistent methods and the table i shows determined values exhibit good agreement . even though exponents are close to mean - field interaction model , they do not exactly match with the values predicted for established universality classes . we have further checked the consistency of the values using magnetic scaling equation . according to this scaling hypothesis , the magnetization @xmath59 , the magnetic field @xmath22 and the temperature @xmath21 obey following relationship,@xcite @xmath60 where the @xmath61 and @xmath62 are the regular functions . 8 implies that renormalized magnetization @xmath63 = @xmath64 plotted as function of renormalized field @xmath65 = @xmath66 with correct set of exponents and @xmath3 will fall in two distinct branches ; one above @xmath3 and another below @xmath3 . the main panel of fig . 8 shows plotting of eq . 8 and the inset shows the same plot in logarithmic scale . it is evident that renormalized @xmath67 distinctly fall in two set of branches . 8 conclusively shows the estimated critical exponents @xmath37 , @xmath38 and @xmath39 and critical temperature @xmath3 are very authentic within the experimental accuracy . to further check the consistency of the determined critical exponents and @xmath3 we have plotted @xmath63 and @xmath65 in form of arrott plot i.e. , @xmath68 vs @xmath69 as shown in fig . 9 . the figure shows isotherms below and above @xmath3 fall into two branches . this rigorous exercise confirms that the obtained critical exponents and @xmath3 are correct . are plotted following eq . 8 for sr@xmath1iro@xmath2 . the isotherms are clearly scaled into two branches below and above @xmath70 with set of exponents @xmath37 and @xmath38 and t@xmath71.,width=302 ] the estimated critical exponents of sr@xmath1iro@xmath2 do not exactly follow any theoretical models , however , the values are quite close to mean - field interaction model . nonetheless , exponent values suggest spin interaction is spatially extended . to further understand the nature of magnetic interaction in this material we have followed a model which suggests that spin interaction decays with spatial distance @xmath72 as ; @xmath73= @xmath74 where @xmath0 is the dimensionality of system and @xmath75 is the range of interaction.@xcite according to this model , the @xmath75 @xmath76 2 implies a faster decay of @xmath73 , hence the spin interaction is of short - range type . on the other hand , @xmath75 @xmath77 1.5 is indicative of long - range type interaction . further , using the renormalization group approach the exponent for magnetic susceptibility @xmath38 can be calculated for a particular value of @xmath78 as following;@xcite @xmath79 \delta\sigma^2 \end{aligned}\end{aligned}\ ] ] where @xmath80 and @xmath81 . following eq . 9 , we have calculated @xmath38 by varying @xmath75 for different set of @xmath78 . for @xmath75 = 1.13 and @xmath78 = @xmath82 , we obtain @xmath38 = 1.145 which turns out to be close to the value experimentally determined for present material ( see table i ) . the @xmath75 = 1.13 ( less than 1.5 ) implies that the nature of spin interaction in sr@xmath1iro@xmath2 is of long - range type thus supporting the estimated exponent values . the @xmath0 = 2 indicates 2-dimensional nature of spin interaction which is quite expected considering the layered structure of this material as realized from generalized ruddlesden - popper series ( sriro@xmath83)@xmath84(sro ) with @xmath9 = 1 . in fact , 2-dimensional spin interaction has been concluded from resonant magnetic x - ray diffuse scattering measurements showing an anisotropic magnetic interaction along in - plane and out - plane direction.@xcite the study has shown in - plane spin correlation survives with sizable strength at temperature much higher ( @xmath76 25 k ) than the magnetic ordering temperature 228.5 k and has given an estimate of exchange coupling constant @xmath85 @xmath14 0.1 ev . on the other hand , out - plane spin correlation exhibits a critical divergence at ordering temperature and shows much weak coupling constant @xmath85 @xmath14 @xmath86ev . the @xmath9 = 3 signifies heisenberg type spin interaction . although this is not consistent with the present set of estimated exponents , however , it is likely that short - range interaction coexist with the long - range one which is discussed in net section . and @xmath65 are plotted in form of @xmath87 vs @xmath69 for sr@xmath1iro@xmath2.,width=302 ] the present material sr@xmath1iro@xmath2 shows weak ferromagnetism arising out of canted type antiferromagnetic ordering which justifies the analysis of critical behavior . in fact , low - frequency ferromagnetic resonance has been evidenced in electron spin resonance ( esr ) study for this material @xcite . with the same spirit , we have attempted to analyze the low temperature thermal demagnetization phenomenon using both spin - wave ( sw ) and single particle ( sp ) model . in the picture of localized moment , the thermal demagnetization at low temperature is generally explained with the spin - wave excitations which follows bloch equation,@xcite @xmath88 where @xmath89 and @xmath90 are the coefficients , @xmath91 is the magnetization at 0 k. the @xmath92 term arises due to harmonic contribution and the @xmath93 term originates from higher order term in spin - wave dispersion relation . the spin - wave stiffness constant @xmath94 can be calculated as , @xmath95^{2/3}\end{aligned}\ ] ] where @xmath96 is the boltzmann constant and @xmath97 is the density of material . in parallel picture of itinerant or band magnetism where the net moment of system is directly proportional to the displacement energy between spin - up and spin - down subbands , the thermal demagnetization is realized as a result of excitation of electrons from one subband to other . the single - particle excitation is generally expressed as,@xcite @xmath98 where @xmath99 is the coefficients , @xmath100 is the energy gap between the top of full sub - band and the fermi level and k@xmath101 is the boltzmann constant . in a system where the decrease of magnetization with temperature is not significant these processes are quite independent . thus , the thermal demagnetization can be explained with the contribution from both spin - wave and single - particle excitation . the lower inset shows fitting of data using spin - wave model ( eq . 10 ) and upper inset shows the same for single - particle model ( eq . the main panel shows fitting incorporating both model ( see text).,width=302 ] fig . 10 shows temperature dependent reduced magnetization @xmath102 which is deduced from @xmath103 data ( fig . the @xmath91 has been estimated from extrapolation of @xmath104 data . initially , we tried to fit the data using eq . 10 with only @xmath92 dependence , however , it did not yield good fitting ( lower inset , fig 10 ) . the addition of @xmath93 term also did not improve much the fitting . similarly , fitting with only sp model using eq.12 also does not give best result , though it is better than sw model . we found , however , best fitting upon incorporating both sw ( up to @xmath92 term in eq . 10 ) and sp model ( eq . the main panel of fig . 10 shows fitting of data till 0.4@xmath3 . the fitting yields @xmath89 = 3.68 @xmath105 10@xmath106 k@xmath107 , @xmath99 = 1.04 @xmath105 10@xmath106 k@xmath108 and @xmath100 = 1.159(2 ) mev . using eq . 11 , we have calculated stiffness constant @xmath94 = 3.365(6 ) ev@xmath109 . the high value of @xmath94 can be understood considering canted type antiferromagnetic ordering in this material which gives very low @xmath91 . the low value of gap @xmath100 = 1.159(2 ) mev is quite interesting . note , that similarly small afm spin excitation gap ( @xmath14 0.85 mev ) has been observed in esr study @xcite . to understand this a detailed investigation is required . nonetheless , this analysis primarily shows presence of both localized and itinerant type of magnetization in sr@xmath1iro@xmath2 which is consistent with the fact that estimated exponents do not exactly match with any particular model ( table i ) . moreover , simultaneous presence of both slater and mott type of insulating phase which is essentially related to itinerant and localized model magnetism , respectively has been previously shown using both experiment and theory for sr@xmath1iro@xmath2 @xcite . iro@xmath2 . different plots correspond to different highest applied magnetic fields . lines are guide to eyes . inset shows linear dependence of change of magnetic entropy with scaled magnetic field @xmath110.,width=302 ] magnetocaloric effect ( mce ) or the change in magnetic entropy ( @xmath24 ) of a material upon application of magnetic field is a vital tool to understand the magnetic phase transition as well as the magnetic state . the change in entropy with field is related with change of magnetization with temperature through thermodynamic maxwell relation as,@xcite iro@xmath2 . inset shows derivative of resistivity as a function of temperature.,width=302 ] @xmath111 the change in magnetic entropy @xmath112 can be calculated as , @xmath113 when magnetization data is at discrete value of temperature and field intervals , @xmath112 can be calculated using following relation , @xmath114 the @xmath24 has been calculated for sr@xmath1iro@xmath2 using eq . 15 from isotherms @xmath25 collected at different temperatures . the @xmath24 has been calculated up to different magnetic fields . the @xmath115 plotted in fig . 11 shows distinct peak around pm to fm phase transition which is also expected from eq . 14 . at low temperature , where @xmath26 shows downfall ( see fig . 1 ) , a weak peak in @xmath115 is observed in fig . 11 . the calculated @xmath24 for sr@xmath1iro@xmath2 is rather low which is justified considering the fact that this material is a weak ferromagnetic system . further , we have used mce to understand the nature of magnetic state as it is shown that in case of mean - field model , @xmath24 shows a power law dependence with scaled filed as , @xmath24 @xmath116 @xmath110.@xcite the inset of fig . 11 , in deed , shows @xmath24 varies linearly with @xmath110 which further confirms magnetic interaction in this material follows mean - field spin interaction model . the estimated critical exponents in table i imply that the magnetic interaction in sr@xmath1iro@xmath2 do not exactly follow the mean - field model but the nature of interaction is quite close to the mean - field type . the analysis of thermal demagnetization ( fig . 10 ) suggests that neither localized spin - wave model nor itinerant stoner single - particle model can fully explain our data . in fact , inclusion of both the model is necessary to understand the thermal demagnetization effect ( fig . 10 ) . following the similar trend , the mce or the change in magnetic entropy ( @xmath24 ) shows a field dependence as , @xmath117 which is in conformity with mean - field model ( inset of fig . these experimental observations indicate that the insulating phase in sr@xmath1iro@xmath2 is more close to slater - type , however , some mott contribution is also present . this coexistence of both these types of mechanism for insulating phase in sr@xmath1iro@xmath2 has also been captured in other theoretical and experimental investigations.@xcite while the previous experimental data have some evidences in favor of slater mechanism the resistivity data , on other hand , have not shown clear metal to insulator transition around magnetic transition till date . in fig . 12 , we have shown resistivity ( @xmath97 ) with temperature for present sr@xmath1iro@xmath2 showing an insulating behavior throughout the temperature range . interestingly , the temperature derivation of resistivity ( @xmath118 ) which is plotted in inset shows distinct change in slope across the magnetic transition temperature @xmath119 or @xmath3 which implies a softening of insulating behavior in pm state . in view of slater model this behavior is not surprising considering the fact that the in - plane exchange interaction in this material is quite robust and survives at temperature much higher than the transition temperature @xmath119,@xcite which probably does not allow the metallic phase in pm state . the simultaneous presence of mott - type insulating phase may also be responsible for continued insulating phase in pm state . in conclusion , in an aim to understand the nature of insulating phase in sr@xmath1iro@xmath2 , whether it is of slater- or mott - type , we have estimated critical exponents around the magnetic phase transition , analyzed the thermal demagnetization data and calculated the change in magnetic entropy . the exponent values do not exactly match the values predicted for universality classes , however , the values are quite close to mean - field model . the analysis of exponent using renormalization group approach suggest spin interaction is of 2-dimensional heisenberg type having extended character . the thermal demagnetization behavior at low temperature can only be explained with inclusion of both localized spin - wave and itinerant stoner single - particle model . the functional change in magnetic entropy with the applied field is suggestive of mean - field model . we conclude that the insulating phase in sr@xmath1iro@xmath2 is close to slater - type though some contribution due to mott - type is also present . we acknowledge ugc - dae csr , indore for magnetization data . we sincerely thank alok banerjee for the magnetization data and discussions . we are also thankful kranti kumar for the help in measurements . b. j. kim , h. jin , s. j. moon , j .- y . park , c. s. leem , j. yu , t.w . noh , c. kim , s .- oh , j .- h . park , v. durairaj , g. cao , and e. rotenberg , phys . rev . * 101 * , 076402 ( 2008 ) . m. k. crawford , m. a. subramanian , r. l. harlow , j. a. fernandez - baca , z. r. wang and d. c. johnston , phys . b * 49 * , 9198 ( 1994 ) . g. cao , y. xin , c. s. alexander , j. e. crow , p. schlottmann , m. k. crawford , r. l. harlow , and w. marshall , phys . b * 66 * , 214412 ( 2002 ) . s. j. moon , h. jin , k.w . kim , w. s. choi , y. s. lee , j. yu , g. cao , a. sumi , h. funakubo , c. bernhard , and t.w . noh1 , phys . * 101 * , 226402 ( 2008 ) . imtiaz noor bhatti , r rawat , a banerjee and a k pramanik , j. phys . : condens . matter * 27 * , 016005 ( 2014 ) . fa wang and t. senthil , phys . lett . * 106 * , 136402 ( 2011 ) . y. yang , w. s. wang j. g. liu , h. chen , j. h. dai , and q. h. wang , phys . b. * 89 * , 094518 ( 2014 ) . h. watanabe , t. shirakawa , and s. yunoki , phys . lett . * 110 * , 027002 ( 2013 ) . y. gao , t. zhou , h. huang and q. h. wang sci . rep . * 5 * , 9251 ( 2015 ) . g. jackeli and g. khaliullin , phys . lett . * 102 * , 017205 ( 2009 ) . b. j. kim , h. ohsumi , t. komesu , s. sakai , t. morita , h. takagi , t. arima , science * 323 * , 1329 ( 2009 ) . f. gebhard , _ the mott metal - insulator transition _ , ( springer - verlag , berlin , ( 1997 ) . j. kim , d. casa , m. h. upton , t. gog , young - june kim , j. f. mitchell , m. van veenendaal , m. daghofer , j. van den brink , g. khaliullin , and b. j. kim , phys 108 * , 177003 ( 2012 ) . d. haskel , g. fabbris , m. zhernenkov , p. p. kong , c. q. jin , g. cao , and m. van veenendaal , phys . lett . * 109 * , 027204 ( 2012 ) . f. ye , s. chi , b. c. chakoumakos , j. a. fernandez - baca , t. qi and g. cao , phys . rev . b * 87 * , 140406 ( 2013 ) . r. arita , j. kunes , a.v . kozhevnikov , a. g. eguiluz , and m. imada , phys . lett . * 108 * , 086403 ( 2012 ) . h. watanabe , tomonori shirakawa , and seiji yunoki , phys . b. * 89 * , 165115 ( 2014 ) . d. hsieh , f. mahmood , d. h. torchinsky , g. cao , and n. gedik , phys . b. * 86 * , 035128 ( 2014 ) . q. li , g. cao , s. okamoto , j. yi , w. lin , b. c. sales , j. yan , r. arita , j. kune@xmath120 , a. v. kozhevnikov , a. g. eguiluz , m. imada , z. gai , m. pan , and d. g. mandrus , sci . rep . * 3 * , 3073 ( 2013 ) . h. e. stanley , _ introduction to phase transitions and critical phenomenon _ , ( oxford university press , new york , ( 1971 ) . a. arrott , phys . rev . * 108 * , 1394 ( 1957 ) . j. s. kouvel and m. e. fisher , phys . rev . * 136 * , a1626 ( 1964 ) . a. arrott and j. e. noakes , phys . * 19 * , 786 ( 1967 ) . s. fujiyama , h. ohsumi , t. komesu , j. matsuno , b. j. kim , m. takata , t. arima , and h. takagi , phys . * 108 * , 247212 ( 2012 ) . b. widom , j. chem . phys . * 43 * , 3898 ( 1965);j . phys . * 41*,1633 ( 1964 ) . m. e. fisher , shang - keng ma , and b. g. nickel , phys . * 29 * , 917 ( 1972 ) . a. k. pramanik and a. banerjee , phys . b. * 79 * , 214426 ( 2009 ) . s. f. fischer , s. n. kaul , and h. kronmller , phys . b. * 65 * , 064443 ( 2002 ) . s. bahr , a. alfonsov , g. jackeli , g. khaliullin , a. matsumoto , t. takayama , t. takagi , b. buchner , and v. kataev _ phys . b. _ * 89 * , 180401 ( 2014 ) . a. das and a. k. majumdar , j. magn . magn . mater . * 128 * , 47 ( 1993 ) . s. n. kaul , phys . b * 27 * , 5761 ( 1993 ) . s. n. kaul , j. phys . : condens . matter * 3 * , 4027 ( 1991 ) . m. h. phana and s. c. yu , j. magn . * 308 * , 325 ( 2007 ) . h. oesterreicher and f. t. parker , j. appl . phys . * 55 * , 4334 ( 1984 ) . q. dong , h. zhang , j. shen , j. sun and b. shen , journal of magn . magn . mater . * 319 * , 56 ( 2007 ) .

the nature of insulating phase in 5@xmath0 based sr@xmath1iro@xmath2 is quite debated as the theoretical as well as experimental investigations have put forward evidences in favor of both magnetically driven slater - type and interaction driven mott - type insulator . to understand this insulating behavior , we have investigated the nature of magnetic state in sr@xmath1iro@xmath2 through studying critical exponents , low temperature thermal demagnetization and magnetocaloric effect . the estimated critical exponents do not exactly match with any universality class , however , the values obey the scaling behavior . the exponent values suggest that spin interaction in present material is close to mean - field model . the analysis of low temperature thermal demagnetization data , however , shows dual presence of localized- and itinerant - type of magnetic interaction . moreover , field dependent change in magnetic entropy indicates magnetic interaction is close to mean - field type . while this material shows an insulating behavior across the magnetic transition , yet a distinct change in slope in resistivity is observed around @xmath3 . we infer that though the insulating phase in sr@xmath1iro@xmath2 more close to be slater - type but the simultaneous presence of both slater- and mott - type is the likely scenario for this material .

recently , kfe@xmath3as@xmath3 ( k-122 ) has attracted special attention due to several unusual physical properties such as a very large quasiparticle effective mass @xcite and superconductivity ( sc ) with nodes in the superconducting gap @xcite . such a large mass enhancement is usually ascribed to the presence of pronounced correlation effects . however , in the present case the nature and the strength of these correlation effects as well as the symmetry of the superconducting order parameter are still under debate . some experiments were interpreted in favor of @xmath7 -wave sc with accidental nodes @xcite whereas other favor @xmath8- wave sc @xcite . based on their theoretical calculations the authors of ref . argue that the electronic correlations are enhanced in the strongly hole - overdoped k-122 as compared to lower hole doping levels in other 122 systems due to the changes in the fermi - surface topology . this might affect the low - energy spin excitation spectra and cause an increasing density of states ( dos ) at the fermi energy @xmath9 . however , this calculation underestimates the correlation effect found experimentally . recently , it was proposed that the unexpected large effective mass enhancement of the order of 9 to 10 could be related to the proximity of an orbital - selective mott transition for k-122 @xcite . the relation between the enhanced correlation effects and the sc remains still unclear . therefore , tuning of the physical properties of k-122 by external or chemical pressure and disorder can provide more insight into the pairing symmetry and the present correlation effects . recently , several studies of the physical properties of k-122 @xcite and its sister compounds cs-122 @xcite , and rb-122 @xcite under external pressure have been reported . tafti _ et al . _ observed that the superconducting critical temperature @xmath0 of k-122 at first monotonically decreases with increasing hydrostatic pressure , then reaches a minimum at @xmath10 1.7 gpa ( we adopt from ref . referring @xmath11 gpa as critical pressure ) , and finally shows a reversal behavior @xcite . the value of the critical pressure is nearly independent of the amount of the disorder in k-122 but it is reduced in the unsubstituted clean cs-122 @xcite . however , v. taufour _ et al . _ , showed that the pressure dependence of @xmath0 for k-122 is sensitive to the pressure medium @xcite , and that for the best pressure ( hydrostatic ) conditions the reversal behavior of @xmath0 is replaced by a broad minimum followed by a nearly pressure independent @xmath0 value above 2.5 gpa . from the almost pressure independent behavior of the hall coefficient extrapolated to @xmath12 = 0 as well as from the residual resistivity , it was concluded that the topology of the fermi surface does not change across the critical pressure.@xcite this point of view was , also , supported by recent de haas - van alphen ( dhva ) measurements under pressure , where at least two small but accessible fermi surface sheets ( fss ) , namely the smallest cylinder @xmath13 around the @xmath14-point and the propeller blade - like fss @xmath15 around the x point of the brillouin zone ( bz ) , do not show a lifshitz transition up to 2.47 gpa.@xcite the evolution of the fermi surface is accompanied by a reduction of the effective quasi - particle masses without any special feature at the critical pressure . the reduction of the effective masses and the decrease of @xmath0 up to the pressure about 2 gpa can be qualitatively explained by a reduction of coupling constants . in order to explain the @xmath0 minimum it was supposed that at the critical pressure a change of the superconducting gap symmetry from @xmath8- wave to @xmath7- wave @xcite or from nodal @xmath7 -wave to a nodeless one @xcite should occur . the fast increase of the inelastic scattering rate at the critical pressure obtained from the transport measurements supports such a hypothesis.@xcite alternatively , v. taufour _ et al . _ proposed that a change of symmetry under pressure may not occur at all.@xcite these authors empirically supposed that the critical pressure could be a consequence of anisotropic pressure derivatives , since the pressure derivatives along the @xmath16 axis and the @xmath17 axis are negative and positive , respectively . in this case , the critical pressure corresponds to the critical pressure in the @xmath18 plane above which the slope of @xmath19 depends on the value of @xmath20 , only , whereas @xmath21 for @xmath22 gpa . these authors have also shown quantitatively that the behavior of the upper critical fields in crossing the critical pressure can be accounted for by @xmath23 modulations of the superconducting gap , only . the amount and type of disorder provide additional possibilities to tune the superconducting properties . recently , the effect of na substitution on the superconducting and the normal - state properties of k-122 was investigated.@xcite it was shown that the temperature dependence of the specific heat of k@xmath1na@xmath2fe@xmath3as@xmath3 single crystals can be explained by multiband @xmath8-wave sc . the disorder induced by the na substitution suppresses the small superconducting gaps and leads to gapless sc with a large residual dos . the specific heat jump scales approximately with a power - law , @xmath24 , with @xmath25 determined by the impurity scattering rate , in contrast to most iron - pnictide superconductors described by the remarkable budko - ni - canfield ( bnc ) scaling @xmath26 . @xcite however , the na substitution might also produce chemical pressure which could additionally affect @xmath0 . in this work , we show that a partial na substitution produces non - hydrostatic chemical pressure with a larger compression along the @xmath17 axis . thereby the netto ( total ) effect of the chemical pressure on @xmath0 is strongly reduced due to anisotropic pressure derivatives with different signs along the @xmath17 and @xmath16(@xmath27 ) axis . therefore , the pressure dependence of @xmath6 ( without pair - breaking ) and the corresponding coupling constant are nearly independent of the na substitution . furthermore , we present a theoretical analysis of the band structure of kfe@xmath3as@xmath3 under pressure within the first - principles density functional theory in the generalized gradient approximation ( gga ) . the calculations indicate that the minimum in the pressure dependence of @xmath0 can be qualitatively explained by a relatively small non - monotonic variation of the density of states ( dos ) at the fermi level . in general , our results suggests that the superconductivity in k-122 is driven by a single leading band . + k@xmath1na@xmath5fe@xmath28as@xmath28 single crystals with a typical mass of about 1 - 2 mg and several mm in - plane dimensions were grown by the self - flux method . the compositions and the phase purity of the investigated samples were determined by an edx analysis in a scanning electron microscope , and by x - ray analysis ; for details see refs . . the two stoichiometric kfe@xmath3as@xmath3 samples were prepared using different fluxes : s1 from a feas flux and s2 from a kas one , for details see refs . @xcite . additionally , we performed single crystal diffraction measurements at room temperature and at @xmath12 = 100 k. the crystallographic parameters were extracted from single crystal diffraction data . the room temperature single crystal x - ray diffraction data were collected on a bruker kappa apex - ii ccd diffractometer using mo - k@xmath13 radiation . @xmath29- and @xmath30-scans were recorded with an increment of 0.3 , integration and corrections for oblique incidence and polarization were performed within saint @xcite . a multiscan absorption correction was applied using sadabs @xcite . structure solutions and refinements were done with the jana2006 package @xcite . single crystal x - ray diffraction analysis at @xmath12 = 100 k was carried out with an ipds ( stoe ) diffractometer using graphite monochromated mo - k@xmath13 radiation . numerical absorption corrections were applied for both data sets with transmission coefficients of @xmath31 = 0.0077/0.3052 and 0.0495/0.06333 , respectively . the structures were solved and anisotropically refined with the shelx program package.@xcite the relative na / k contents were included into the refinements as an independent parameter ( x ) . the temperature dependence of the electrical resistivity was measured by the standard four - contact method in the quantum design physical property measurement system ( ppms ) . the temperature dependent , zero - field - cooled ( zfc ) magnetization under pressure was measured in a quantum design mpms - squid magnetometer down to 1.8 k in a magnetic field of 10 oe . the samples position was not perfectly fixed during the magnetization measurements under pressure . therefore , in some cases samples could be tilted with respect to the applied field direction after the application of the pressure . this affects the amplitude of the superconducting diamagnetic signal ( due to a variation in the demagnetization factor ) as can be seen in fig . [ mag_all ] . however , the defined @xmath0 values are not affected considerably , since the measurements were performed at low applied magnetic field as compared to the critical fields of the investigated samples . for the dc magnetization measurement under pressure , we used a commercial pressure cell easylab mcell 10 tailored for a mpms of quantum design for the pressure measurements up to 1 gpa and a homemade pressure cell ( hpc ) for the pressure measurements up to 3 gpa . the hpc has been constructed for a commercial mpms . its cell design is mostly related to a diamond anvil cell ( dac ) , in which two opposing cone - shaped ceramic anvils compress a gasket that serves as the sample chamber.@xcite a supplied uniaxial pressure will be transformed into uniform hydrostatic pressure by using daphne7373 oil as transmitting medium . the cell can be equipped with different types of sample chambers , for which the sample space varied between either 0.7 mm in diameter and 0.4 mm in depth , or on the other hand 0.3 mm in diameter and 0.2 mm in depth . using these different chambers a maximum hydrostatic pressure up to 3 gpa and 6 gpa can be generated , respectively . each mechanical part of the pressure cell , except for the anvils , is made from a very low - magnetic cube alloy , which allows measurements at a weak background signal of the cell itself . @xcite additionally , the magnetization data have been corrected for the contribution of the background signal by subtracting the extrapolated ( below @xmath0 ) background signal obtained at @xmath32 . the pressure was determined at low temperatures using superconducting sn in case of the quantum design pressure cell and pb in case of the homemade cell as a low - temperature pressure gauge . for most of the measurements the pressure dependence of @xmath0 was linear up to 1.5 gpa ( see fig . [ tc_p ] ) . however , using the easylab mcell 10 pressure cell in some cases we observed deviations from a linear @xmath19 dependence at pressures below 0.1 gpa . to verify the origin of this behavior at low pressures , we performed additional measurements in a he - gas cube pressure cell up to the maximum pressure of 0.5 gpa . in the experiment a cube cell connected to a he - gas compressor was used for finite @xmath33 measurements . @xcite an in sample was used for an in - situ determination of the pressure . @xcite the use of 4-he as a pressure - transmitting medium ensures truly hydrostatic pressure conditions at low temperatures . @xcite the susceptibility data have been corrected for the contribution of the sample holder , including the pressure cell , which was determined independently . in the case of the experiments with he as a pressure medium the non - linear behavior in the @xmath19 dependence at low pressures was not observed . therefore , we may conclude that the observed deviation from the linear @xmath19 dependence is related to deviation from the hydrostatic conditions . in this work , we consider the data with a linear @xmath19 dependence at low pressure , only , assuming that in this case the measurements where performed under hydrostatic pressure conditions . + dependence of the resistivity of k@xmath1na@xmath2fe@xmath28as@xmath28 at zero pressure . the data for @xmath34 = 0.18 and @xmath34 = 0.31 are taken from the supplementary information in ref . . ( a ) the temperature range below 10 k , symbols - experimental data , solid lines - single band fit eq.[eq1 ] , dashed lines - multiband fit eq . ( b ) the temperature range up to 300 k. inset : the dependence of the exponent @xmath35 vs. the residual resistivity @xmath36 , see eq . the data for k@xmath1na@xmath2fe@xmath28as@xmath28 crystals ( @xmath37 ) are taken from the supplementary information in ref . , for @xmath34 = 0 ( open rhombus ) from ref . . the solid line is a guide - to - the - eye . ] fig.[res ] shows the @xmath12 dependence of the electrical resistivity @xmath38 of k@xmath1na@xmath2fe@xmath28as@xmath28 single crystals . to remove the uncertainty in the absolute value of @xmath38 between our transport data and the published one in literature associated with geometric factors , we used a normalization procedure as proposed in ref . . in this way , the resistivity data @xmath38 of each sample were multiplied by a factor @xmath39k@xmath40k@xmath41 , where @xmath42k@xmath41 is the measured resistivity of the sample at @xmath12 = 300 k and @xmath39k@xmath43 = 0.36 m@xmath44 cm is our average resistivity value over all of our samples at @xmath12 = 300 k. below we show only normalized @xmath38 data . to fit the data below 15 k , we used the standard single - band expression @xmath45 the obtained residual resistivity @xmath46 monotonically increases with the na doping level as shown in fig.[res]a . in turn , @xmath0 decreases with the increasing @xmath46 in accord with the presence of the disorder induced pair - breaking.@xcite we have found that the exponent @xmath35 is always below 2 : @xmath47 for our stoichiometric k-122 samples and it decreases to @xmath48 for the single crystals with a high na substitution level . in the case of single - band materials such a deviation from a @xmath4 dependence would signal a non - fermi liquid ( nfl ) behavior . for example , the deviation from the fl behavior in the strongly disordered k-122 has been interpreted as evidence for quantum criticality @xcite . however , in a multiband system , which k-122 is , the deviation from the @xmath4 behavior can be observed even in the case when all @xmath49 bands behave like @xmath50 . for the resulting resistivity of the multiband system we have : @xmath51^{-1}. \label{eq2}\ ] ] strictly speaking , in the case of k-122 a four ( five ) band - model should be considered , if one neglects ( takes into account ) the small 3-dimensional band near the z point of the bz ( see fig.[fs_exp ] ) with 8(10 ) independent parameters . to demonstrate that the data can be fitted by eq.[eq2 ] we considered for the sake of simplicity an effective two - band model as shown in fig.[res]a . ( the details of the fitting parameters are given in table [ tab:2 ] . ) as can be seen , a two - band analysis results in quite different @xmath52 values . the value of the coefficient @xmath53 is close to the values reported for k-122 compounds obtained using the single - band eq . [ eq1 ] . @xcite . however , the value of the coefficient @xmath54 is about one order of magnitude larger than that of @xmath53 . this difference can be understood if we consider that @xmath55 , where @xmath35 is the density of the carriers , @xmath56 is the bare density of states ( dos ) at the fermi energy , @xmath57 is the bare fermi velocity in the @xmath18 plane , and @xmath58 is the frequency which determines the magnitude of the frequency dependent term in the imaginary part of the self - energy and contains all of the many - body effects . @xcite then , the ratio @xmath59 can be explained by the ratio of @xmath60 , only , by taking into account rather different band dependent quasiparticle effective masses for k-122 @xcite . therefore , the obtained parameters ( see table [ tab:2 ] ) have reasonable values for the k@xmath1na@xmath2fe@xmath28as@xmath28 system under consideration which suggests that the nominally nfl behavior of the total @xmath38 can be explained by a multiband effect with all bands in a fl state . fig.[mag_all ] shows the @xmath12 dependence of the magnetization for k@xmath1na@xmath2fe@xmath28as@xmath28 single crystals . the @xmath0 for the samples is defined using linear extrapolations as shown in fig . [ mag_all ] . using this criterion for the critical temperature similar @xmath0 values are obtained from our magnetization data and from specific heat measurements performed on the crystals from the same batch @xcite . this @xmath0 criterion results in slightly lower @xmath0 values than that obtained from the zero - resistivity value usually used as a criterion for @xmath0 in the transport measurements under pressure.@xcite however , the observed small discrepancy between these two criteria does not affect the main conclusion of our work . it is seen in fig.[mag_all ] that @xmath0 of all samples monotonically decreases with pressure up to about 2 gpa . the sample s2 ( @xmath34 = 0 ) was measured up to the maximum pressure of about 3 gpa fig.[mag_all]b . it is seen in fig.[tc_p]a that the pressure dependence of @xmath0 for this sample has a shallow minimum at about 2 gpa in accord with the data obtained in the experiments with he as a pressure medium.@xcite the pressure dependence of @xmath0 for our k@xmath1na@xmath2fe@xmath28as@xmath28 system and the available data in the literature for k-122,@xcite cs-122 @xcite and rb-122 @xcite are summarized in fig.[tc_p]a . one can see that k@xmath1na@xmath5fe@xmath28as@xmath28 and stoichiometric k-122 with different @xmath0 values at zero pressure exhibit a similar pressure derivative @xmath61 -1 k / gpa . however , the sister compounds rb-122 and cs-122 show clearly different pressure derivatives -1.35 k / gpa and -0.85 k / gpa , correspondingly . as was shown in our previous work , the @xmath0 of the k@xmath1na@xmath2fe@xmath28as@xmath28 system at zero pressure is defined mainly by a single strongly coupled superconducting band . @xcite to demonstrate that the na substitution suppresses @xmath0 due to the pair - breaking effect , only , and that it does not affect the superconducting coupling constants @xmath62 we analyzed our experimental data adopting the single - band abrikosov - gorkov ( ag ) formula modified for the @xmath8-wave case @xcite @xmath63 where @xmath64 is the pair - breaking parameter and @xmath65 is the effective scattering rate due to impurities created by the na substitution . the quantities @xmath66 , and @xmath6 can be pressure dependent . taking this into account , we have found that all available data for the k@xmath1na@xmath5fe@xmath28as@xmath28 system and the disordered k-122 crystals can be fitted by the ag formula assuming that the critical temperature of the clean system ( without pair - breaking impurities ) varies linearly with pressure and can be approximated by @xmath67 k @xmath68 up to @xmath69 1.5 gpa . in contrast , it is found that @xmath70 is pressure independent . in general , the observed ag behavior in the k@xmath1na@xmath5fe@xmath28as@xmath28 system may be considered as a justification that na acts as a pair - breaking impurity in agreement with ref . . the independence of @xmath71 on the na concentration is illustrated in fig . [ tc_p]b , where we plot the normalized @xmath72 versus the residual resistivity @xmath46 measured at zero pressure divided by @xmath71 . it is seen that all available data up to 1.5 gpa scale on a single curve . the observed scaling indicates , also , that both @xmath70 and @xmath73 is nearly pressure independent . the insensitivity of @xmath46 to pressure for k-122 is confirmed experimentally . @xcite interestingly , the cs-122 data can be scaled to the same curve using their different @xmath74 , only . note , that the clean limit @xmath75 k is rather reasonable for cs-122 compound taking into account available literature data.@xcite . for example , the reported in ref . @xmath76 2.6 k is quite close to our adopted @xmath77 value . + for k@xmath1na@xmath2fe@xmath28as@xmath28 samples from this work and those taken from the literature for stoichiometric k-122 @xcite , cs122 @xcite , rb122 @xcite . dashed lines : linear fit and solid lines : results obtained from eq . b ) the normalized critical temperature @xmath72 vs. the ratio between the residual resistivity at zero pressure and the critical temperature for clean k-122 @xmath78 . the solid line denotes the abrikosov - gorkov formula . the @xmath46 values for disordered k-122 and cs-122 samples ( in brackets ) are taken from ref . and have been normalized according to procedure described in the text . ] na@xmath2fe@xmath28as@xmath28 correspond to the lattice parameters which have been obtained from x - ray data at atmospheric pressure . the data points are shifted along the pressure axis to fit the values for kfe@xmath28as@xmath28 at external hydrostatic pressure . a ) the data obtained at room temperature . close symbols - the experimental data for k@xmath1na@xmath2fe@xmath28as@xmath28 , open symbols are taken from ref . @xcite b ) close symbols - the experimental data for k@xmath1na@xmath2fe@xmath28as@xmath28 obtained at @xmath79 100 k. solid lines - theoretical calculations corresponding to the low - temperature regime . inset : single crystalline diffraction pattern of ( hk0 ) , and ( h0l ) planes for a na doped crystal . ] it is well known that isovalent chemical substitution in iron pnictides suppresses the sdw phase and can induce superconductivity ( see for example refs . ) . in this case one of the relevant factors which controls the phase diagram is the change of the structure parameters @xcite . this effect can be considered as a chemical pressure . to understand the effect of the na substitution on the lattice parameters , we performed single crystal diffraction measurements at room temperature and at @xmath80 k. selected crystallographic results are given in table [ tab:1 ] . the lattice parameters of k@xmath1na@xmath2fe@xmath28as@xmath28 single crystals obtained at @xmath12 = 300 k are compared with the experimental data for k-122 @xcite obtained at room temperature ( fig.[c_structure]a ) and the lattice parameters obtained at @xmath12 = 100 k are compared with the results of our dft calculations relevant for low temperature ( fig.[c_structure]b ) . we note that the lattice parameters only weakly depend on temperature below @xmath81 100 k @xcite . it is seen in fig.[c_structure ] that the chemical pressure due to the na substitution produces a stronger effect along the @xmath17 axis by about a factor of 2 - 3 than in the @xmath18 plane as compared to the hydrostatic external pressure effect on k-122 . taking into account that @xmath0 of k-122 is equally sensitive to uniaxial pressure applied in the @xmath18 plane and along the @xmath17 axis ( @xmath82 ) with the resulting superposed pressure derivative @xmath83 @xcite , we can conclude that the chemical pressure effect on @xmath0 due to the na substitution is rather weak . @xcite this provides additional support for our assumption that the main reason of the observed @xmath0 suppression in na doped crystals is given by the pair - breaking effect . to get a better understanding of the pressure effect on @xmath0 we performed a density functional theory ( dft ) based theoretical analysis of the band structure under pressure . + the calculations were performed using the vienna ab - initio simulation package ( vasp)@xcite within the generalized gradient approximation ( gga)@xcite . the perdew , burke and ernzerhof ( pbe)@xcite functional was used to calculate the exchange - correlation potential . we used projected augmented - wave ( paw ) pseudopotentials@xcite for all the atomic species involved , and in order to achieve a satisfactory degree of convergence the integrations over the brillouin zone ( bz ) was performed considering 12@xmath8412@xmath848 and 24@xmath8424@xmath8424 uniform monkhorst and pack@xcite grids for the conventional ( 4 k , 8 fe and 8 as atoms ) and single formula unit ( 1 k , 2 fe and 2 as atoms ) cells , respectively , and energy cutoff up to 500 ev . the fermi surface cuts of chosen lattice structures ( see below ) as a function of @xmath23 have been computed using a @xmath85 k - points grid . + we studied the influence of the structural properties on the electronic paramagnetic ( pm ) band structure of kfe@xmath3as@xmath3 . in fig . [ bandstr ] , we report the calculated band structure considering the experimental @xcite and the optimized lattice constants in the pm phase . the predicted ( theoretical ) and experimental lattice constants are compared in table [ tab:3 ] . in spite of relatively small differences between experimental and theoretical lattice parameters ( in a range of 1 - 2 % ) , the fermi surface cross - sections calculated with experimental lattice parameters reveal important differences with respect to the theoretical calculated ones . the radii of the @xmath14-centered hole fermi surfaces , namely @xmath13 , @xmath86 and @xmath87 , do not change sensibly , while the small electron - pockets around the x point shrink considering the theoretical lattice constants . the small ( but irrelevant ) electron pocket at z is found to become larger using the theoretical lattice parameters . since the discrepancy between theoretical and experimental crystal lattices affects bands , in the further analysis we stick to the experimental structure @xcite : we consider this one as the reference geometry at @xmath88 , _ i. e. _ , @xmath89 , @xmath90 , and @xmath91 . then , we extracted the lattice parameters calculating the compressibility for the @xmath92 , @xmath93 and @xmath94 by first principles as a function of the pressure . the lattice constants , obtained in this way , are given in tab . [ tab:1a ] . these data was also used for the analysis of the chemical pressure ( see fig . [ c_structure]b ) . the calculated band structure and fermi surface cross - sections with these lattice parameters are given in figs . [ bands_pressure_pm ] , and [ fs_exp ] for different pressures . we underline that experiments found the two outermost fss around the @xmath14 point not intersecting@xcite , however , dft - gga find them crossing . keeping into account this effect on the band structure , the orbital character of the fermi surface is the following : _ i ) _ the main contribution to the @xmath13 fs is from @xmath95 ( along the @xmath14z direction ) and from @xmath96 ( along the @xmath14x direction ) orbitals ; _ ii ) _ we call @xmath86 the band with character @xmath97 ( the green fs along the @xmath14z direction which becomes red along the @xmath14x , see fig . [ fs_exp ] ) ; _ iii ) _ in the same way , we refer to @xmath87 the fs with @xmath96 character ( the outermost red fs along the @xmath14z direction becoming the middle green one along @xmath14x , see fig . [ fs_exp ] ) ; _ iv ) _ the main contribution to the lobes @xmath15 near the x point is from @xmath96 and @xmath97 ; _ v ) _ the character of the three dimensional fs at the z point ( orange band in fig . [ fs_exp ] ) is @xmath97 . as@xmath3 band structure considering different crystal lattice parameters ( namely , the experimental and pm ones , see table [ tab:3 ] ) . ] as@xmath3 band structure for the experimental lattice structure ( taken as @xmath98 gpa geometry ) and lattice parameters extrapolated with compressibility at 1.66 and 3.00 gpa ( see tab . [ tab:1a ] ) . ] in order to understand the origin of the reversed behavior of @xmath0 under pressure , first , we compared the band structure at ambient ( @xmath33 = 0 gpa ) , close to the critical pressure ( @xmath33 = 1.66 gpa ) and at high pressure ( @xmath33 = 3 gpa ) , shown in fig . [ bands_pressure_pm ] . the shape of the relevant fermi surfaces are not strongly affected in this range of pressure , indicating that the origin of the non - monotonic pressure dependence of @xmath0 is not due to changes of the fermi surface topology in accord with experimental findings . @xcite [ cols="^,^ " , ] in this context , the observed ag - type scaling of the pressure dependencies of @xmath0 for nodal k , na-122 and cs-122 suggests that these compounds have the same superconducting gap symmetry . however , the cs-122 and the ( k , na)-122 systems have different @xmath6 and hence different coupling constants . in the case of extended @xmath99-wave pairing symmetry such as @xmath100 -wave sc with accidental nodes , the nodal position is not fixed on the fermi surface @xcite . therefore , nodes can be lifted by impurities , pressure or isovalent substitution . in some cases a reentrant behavior of the nodes can be observed.@xcite in contrast , for k , na-122 single crystals we observed a monotonic increase of the residual electronic specific heat @xmath101 with na substitution and a nearly na independent @xmath4 behavior of @xmath102 at low - temperature . @xcite therefore , our data are consistent with a symmetry protected @xmath8-wave sc in these compounds which is stable at least up to the pressure where @xmath0 has a minimum ( see fig . [ tc_p]b ) . finally , we compare the calculated fermi surface cross sections with the experimental ones obtained from dhva measurements @xcite . these dhva experiments show , that the size of the lobes near the x - point decreases with pressure and that of the @xmath13 cylinder weakly increases up to 2.47 gpa , the maximum pressure reached experimentally . the trend in the calculated fermi surface cross sections for this pressure range , i.e , up to 2.5 gpa ( see fig . [ fs_exp ] ) is consistent with the corresponding dhva frequencies . however , increasing the pressure above 2.5 gpa , e.g. in the case of 3 gpa shown in fig . [ fs_exp](d ) , the size of the lobes maintain their dimension . thus , this prediction could be verified experimentally by studies around this pressure . for this purpose , in tab . [ tab:4 ] we collected the calculated cross section of various fs bands , dhva frequencies and the @xmath103 ratios as found in the literature from both measurements and simulations . together with them , we add our gga results calculated for the experimental crystal structure . the dhva frequencies and the @xmath103-ratio have been estimated calculating the extremal fs areas and the relative orbit frequencies using its implementation in the skeaf code@xcite . the gga calculated fs cross sections are in qualitative agreement with respect to the experiments ( except @xmath15 fs ) . however , the theoretical @xmath103 ratios essentially underestimate experimental values pointing to sizable correlation effects in accord with a previous study . in this context we note that the observed large difference of the effective resistivity coefficients @xmath53 and @xmath54 discussed above ( see tab.[tab:2 ] ) , is likely to be caused by the high - energy mass renormalizations and not related directly to the pairing interactions . this suggestion is also consistent with the significant pressure dependence of the approximate effective single @xmath104 coefficient reported by v. taufour _ et al . _ @xcite . at this point , we extended the study to high pressure and we report the results of calculated dhva frequencies and we compared them with experimental measurements.@xcite the results obtained in the calculations are shown in tab . [ tab:5 ] for the fs s shown in fig . [ fs_exp ] . we observed that the largest dhva frequency increases for the @xmath13-band as a function of the pressure , in agreement with experiment.@xcite at the same time , the lowest frequency slightly decreases with pressure overestimating the increase of the three dimensionality of the electronic structure under pressure as compared to experiment.@xcite we confirm , moreover , the absence of a lifshitz transition under pressure . @xcite note , that the variation of the tdos at the fermi level in the pressure range up to 3 gpa is below 15% as shown in fig . such a relatively small variation of the tdos for multiband systems can be hardly distinguished in experiments such as transport measurements . therefore , we suppose that the observation of a nearly pressure independent hall coefficient and residual resistivity @xcite can not exclude such relatively small variations of the dos . + in conclusion , we investigated the effect of hydrostatic pressure and na substitution on the superconducting and the normal - state properties of kfe@xmath3as@xmath3 . we have found that the na substitution noticeably affects the lattice parameters which can be considered as a non - hydrostatic chemical pressure . however , the comparison of the external hydrostatic pressure with the chemical pressure shows that the total effect of the structural changes on @xmath0 due to na substitution is rather weak . therefore , the main mechanism of the @xmath0 suppression in na doped crystals is a weak pair - breaking effect caused by disorder . the na substitution enhances the residual resistivity and leads to a formal non - fermi liquid behavior in the temperature dependence of the resistivity . we have shown that this deviation from the @xmath4 behavior can be described by multiband effects due to different effective quasi - particle masses for different bands . the pressure dependencies of @xmath0 for the k@xmath1na@xmath2fe@xmath3as@xmath3 system can be described by the single - band abrikosov - gorkov - type scaling curve using the na independent critical temperature in a clean limit @xmath6(p ) . these observations additionally confirm that the @xmath0 suppression for na substitution is mainly caused by the pair - breaking effect . the pressure dependence of @xmath0 for csfe@xmath3as@xmath3 can be also scaled on the same curve , however , with a different @xmath6(p ) . these results suggest the same pairing symmetry for this compound as for the k@xmath1na@xmath2fe@xmath3as@xmath3 system . additionally , we performed theoretical investigations of the band structure of kfe@xmath3as@xmath3 under pressure within the generalized gradient approximation . our results suggest , that the observed pressure dependence of @xmath0 can be explained by the non - monotonic variation of the pdos of the @xmath105 derived band under pressure without a change of the pairing symmetry . this @xmath105 derived band is supposed to be most relevant for the superconductivity of the k@xmath1na@xmath2fe@xmath3as@xmath3 system . this work was supported by the eu - japan project ( no . 283204 super - iron ) , and the dfg through the spp 1458 , the e .- noether program ( wu 595/3 - 1 ( s.w . ) ) . thanks the bmbf for support in the frame of the era.net rus project fesuco no . 245 , e.a . thanks the dfg - grk1621 and i.m . thanks funding from rfbr ( 12 - 03 - 01143-a ) . we acknowledge d. evtushinsky , k. iida , k. nenkov , c. hess , and a. yaresko for fruitful discussion . t. terashima , k. kihou , k. sugii , n. kikugawa , t. matsumoto , s. ishida , c .- h . lee , a. iyo , h. eisaki , and s. uji , phys . b * 89 * , 134520 ( 2014 ) . tafti , a. juneau - fecteau , m.- . delage , s. ren de cotret , j .- reid , a.f . wang , x .- luo , x.h . chen , n. doiron - leyraud , and l. taillefer , nature physics * 9 * , 349 ( 2013 ) . t. terashima , n. kurita , m. kimata , m. tomita , s. tsuchiya , m. imai , a. sato , k. kihou , c .- h . lee , h. kito , h. eisaki , a. iyo , t. saito , h. fukazawa , y. kohori , h. harima , and s. uji , phys . rev . b * 87 * , 224512 ( 2013 ) . f. hardy , a.e . bhmer , d. aoki , p. burger , t. wolf , p. schweiss , r. heid , p. adelmann , y.x . yao , g. kotliar , j. schmalian , and c. meingast , phys . 111 * 027002 ( 2013 ) . v. grinenko , d.v . efremov , s .- l . drechsler , s. aswartham , d. gruner , m. roslova , i. morozov , k. nenkov , s. wurmehl , a.u.b . wolter , b. holzapfel , and b. bchner phys . b * 89 * , 060504(r ) ( 2014 ) . h. fukazawa , t. saito , y. yamada , k. kondo , m. hirano , y. kohori , k. kuga , a. sakai , y. matsumoto , s. nakatsuji , k. kihou , a. iyo , c.h . lee , and h. eisaki , j. phys . jpn . * 80 * , sa118 ( 2011 ) . m. abdel - hafiez , v. grinenko , s. aswartham , i. morozov , m. roslova , o. vakaliuk , s. johnston , d.v . efremov , j. van den brink , h. rosner , m. kumar , c. hess , s. wurmehl , a.u.b . wolter , b. bchner , e.l . green , j. wosnitza , p. vogt , a. reifenberger , c. enss , m. hempel , r. klingeler , and s .- l . drechsler , phys . b * 87 * , 180507(r ) ( 2013 ) . reid , m.a . tanatar , a. juneau - fecteau , r.t . gordon , s.r . de cotret , n. doiron - leyraud , t. saito , h. fukazawa , y. kohori , k. kihou , c.h . lee , a. iyo , h. eisaki , r. prozorov , and l. taillefer , phys . * 109 * , 087001 ( 2012 ) . k. okazaki y. ota , y. kotani , w. malaeb , y. ishida , t. shimojima , t. kiss , s. watanabe , c .- t . chen , k. kihou , c.h . lee , a. iyo , h. eisaki , t. saito , h. fukazawa , y. kohori , k. hashimoto , t. shibauchi , y. matsuda , h. ikeda , h. miyahara , r. arita , a. chainani , s. shin , science * 337 * , 1314 ( 2012 ) . z.p . yin , k. haule and g. kotliar , nature materials * 10 * , 932 ( 2011 ) . f. hardy , r. eder , m. jackson , d. aoki , c. paulsen , t. wolf , p. burger , a. bhmer , p. schweiss , p. adelmann , r.a . fisher , and c. meingasta , j. phys . . jpn . * 83 * , 014711 ( 2014 ) . s. kittaka , y. aoki , n. kase , t. sakakibara , t. saito , h. fukazawa , y. kohori , k. kihou , c .- h . lee , a. iyo , h. eisaki , k. deguchi , n.k . sato , y. tsutsumi , and k. machida , j. phys . soc . jpn . * 83 * 013704 ( 2014 ) . d. watanabe , t. yamashita , y. kawamoto , s. kurata , y. mizukami , t. ohta , s. kasahara , m. yamashita , t. saito , h. fukazawa , y. kohori , s. ishida , k. kihou , c.h . lee , a. iyo , h. eisaki , a.b . vorontsov , t. shibauchi , y. matsuda , phys . b * 89 * , 115112 ( 2014 ) . k. hashimoto , a. serafin , s. tonegawa , r. katsumata , r. okazaki , t. saito , h. fukazawa , y. kohori , k. kihou , c.h . lee , a. iyo , h. eisaki , h. ikeda , y. matsuda , a. carrington , and t. shibauchi , phys . b * 82 * , 014526 ( 2010 ) . h. kim , m. a. tanatar , yong liu , zachary cole sims , chenglin zhang , pengcheng dai , t. a. lograsso , r. prozorov , phys . b * 89 * , 174519 ( 2014 ) . h. ikeda , r. arita , and j. kunes , phys . b * 82 * , 024508 ( 2010 ) . budko , y. liu , t. a. lograsso , and p. c. canfield , phys . b * 86 * , 224514 ( 2012 ) . budko , m. sturza , d.y . chung , m.g . kanatzidis , and p.c . canfield , phys . b * 87 * , 100509(r ) ( 2013 ) . v. taufour , n. foroozani , j.k . lim , m.a . tanatar , u. kaluarachchi , s.k kim , y. liu , t.a . lograsso , v.g . kogan , r. prozorov , s.l . budko , j.s . schilling , p.c . canfield phys . * b * 89 , 220509(r ) ( 2014 ) . p. burger , f. hardy , d. aoki , a.e . bhmer , r. eder , r. heid , t. wolf , p. schweiss , r. fromknecht , m.j . jackson , c. paulsen , and c. meingast , phys . b * 88 * , 014517 ( 2013 ) . tafti , j.p . clancy , m. lapointe - major , c. collignon , s. faucher , j. sears , a. juneau - fecteau , n. doiron - leyraud , a.f . wang , x.g . luo , s. desgreniers , young - june kim , x.h . chen , and l. taillefer , phys . b * 89 * , 134502 ( 2014 ) . k. sasmal , b. lv , b. lorenz , a. m. guloy , f. chen , y .- y . xue , and c .- w . chu , prl * 101 * , 107007 ( 2008 ) . wang , s.y . zhou , x.g . luo , x.c . hong , y.j . yan , j.j . ying , p. cheng , g.j . ye , z.j . xiang , s.y . li , and x. h. chen , phys . rev . b * 89 * , 064510 ( 2014 ) . z. bukowski , s. weyeneth , r. puzniak , j. karpinski , b. batlogg physica c * 470 * , s328 ( 2010 ) . z. shermadini , j. kanter , c. baines , m. bendele , z. bukowski , r. khasanov , h .- h . klauss , h. luetkens , h. maeter , g. pascua , b. batlogg , and a. amato physical review b * 82 * , 144527 ( 2010 ) . z. shermadini , h. luetkens , a. maisuradze , r. khasanov , z. bukowski , h .- h . klauss , and a. amato , phys . b * 86 * , 174516 ( 2012 ) . z. zhang , a. f. wang , x. c. hong , j. zhang , b. y. pan , j. pan , y. xu , x. g. luo , x. h. chen , and s. y. li , arxiv:1403.0191 ( 2014 ) . tafti ; private communication . up to 3 gpa the residual resistivity remains almost constant within deviations of about 5 % . budko , ni ni , and p.c . canfield , phys . b * 79 * , 220516r ( 2009 ) . m. abdel - hafiez , s. aswartham , s. wurmehl , v. grinenko , c. hess , s .- l . drechsler , s. johnston , a.u.b . wolter , and b. bchner , h. rosner , l. boeri , phys . b * 85 * , 134533 ( 2012 ) . v. grinenko , s .- l . drechsler , m. abdel - hafiez , s. aswartham , a.u.b . wolter , s. wurmehl , c. hess , k. nenkov , g. fuchs , d.v . efremov , b. holzapfel , j. van den brink , and b. bchner , phys . status solidi b * 250 * , 593 ( 2013 ) . saint+ ( v. 2011 ) , bruker axs inc . m. sheldrick , sadabs ( version 2008/1 ) , bruker axs inc . v. petricek , m. dusek , and l. palatinus , jana2006 . the crystallographic computing system . institute of physics , prague , czech republic ( 2006 ) . g. m. sheldrick , acta crystallogr . a , * 64 * , 112 ( 2008 ) . p. l. alireza and g. g. lonzraich , rev . * 80 * , 023906 ( 2009 ) . n. tateiwa , y. haga , t. d. matsuda , and z. fisk , rev . * 83 * , 053906 ( 2012 ) . n. tateiwa , y. haga , t. d. matsuda , z. fisk , s. ikeda and h. kobayashi , rev . instrum . * 84 * , 046105 ( 2013 ) . details about the pressure cell construction will be published elsewhere . institute of high - pressure physics , polish academy of sciences , unipress equipment division . l.d . jennings and c.a . swenson , phys . 112 , 31 ( 1958 ) . e. gati , s. khler , d. guterding , b. wolf , s. knner , s. ran , s.l . budko , p.c . canfield , and m. lang , phys . b * 86 * , 220511(r ) ( 2012 ) . dong , s.y . zhou , t.y . guan , h. zhang , y.f . dai , x. qiu , x.f . wang , y. he , x.h . chen , and s.y . li , phys.rev . 104 * , 087005 ( 2010 ) . a.c . jacko , j.o . fjaerestad , b.j . powell , nature phys . * 5 * , 422 ( 2009 ) . radtke , k. levin , h.b . schuttler , m. r. norman , phys . b * 48 * , 653 ( 1993 ) . p. walmsley , c. putzke , l. malone , i. guillamon , d. vignolles , c. proust , s. badoux , a.i . coldea , m.d . watson , s. kasahara , y. mizukami , t. shibauchi , y. matsuda , and a. carrington , phys . * 110 * , 257002(2013 ) . f. rullier - albenque , d. colson , a. forget , p. thuery , and s. poissonnet , phys . b * 81 * , 224503 ( 2010 ) . j. engelmann , v. grinenko , p. chekhonin , w. skrotzki , d.v . efremov , s. oswald , k. iida , r. hhne , j. hnisch , m. hoffmann , f. kurth , l. schultz , and b. holzapfel , nature communications * 4 * ( 2013 ) 2877 . s. avci , o. chmaissem , d.y . chung , s. rosenkranz , e.a . goremychkin , j.p . castellan , i.s . todorov , j.a . schlueter , h. claus , a. daoud - aladine , d.d . khalyavin , m.g . kanatzidis , and r. osborn , phys . b * 85 * , 184507 ( 2012 ) . the uniaxial pressure derivatives were obtained using the ehrenfest relation , i.e. , in the limit of zero external pressure . @xcite in the case of na substituted samples the effect of chemical pressure is rather strong ( about 0.5 - 1.5 gpa ) . in such high pressures the ratio between pressure derivatives may be different from the estimated one using the ehrenfest relation . however , taking into account that @xmath19 of all ( k , na)-122 samples varies nearly linear up to 1.5 gpa and @xmath71 is independent on the na substitution ( see section iii.a ) , we concluded that the ratio between the pressure derivatives is nearly unchanged up to 1.5 gpa . g. kresse and j. furthmuller , phys . b * 54 * , 11169 ( 1996 ) . g. kresse and j. furthmuller , comput . * 6 * , 15 ( 1996 ) . g. kresse and d. joubert , phys . b * 59 * , 1758 ( 1999 ) . perdew , k. burke , and m. ernzerhof , phys . lett . * 77 * , 3865 ( 1996 ) . blchl , phys . b * 50 * , 17953 ( 1994 ) . monkhorst and j.d . pack , phys . b * 13 * , 5188 ( 1976 ) . v. mishra , g. boyd , s. graser , t. maier , p.j . hirschfeld , and d.j . scalapino , phys . b * 79 * , 094512 ( 2009 ) . y. wang , a. kreisel , p.j . hirschfeld , and v. mishra , phys . b * 87 * , 094504 ( 2013 ) . rourke , s.r . julian , comp . comm . * 183 * , 324 ( 2012 ) . t. yoshida , s. ideta , i. nishi , a. fujimori , m. yi , r.g . moore , s.k . mo , d .- h . lu , z .- x . shen , z. hussain , k. kihou , p.m. shirage , h. kito , c.h . lee , a. iyo , h. eisaki , and h. harima , front . phys . * 2 * , 17 ( 2014 ) . t. sato , k. nakayama , y. sekiba , p. richard , y .- m . xu , s. souma , t. takahashi , g.f . chen , j.l . luo , n.l . wang , and h. ding , phys . lett . * 103 * , 047002 ( 2009 ) . zocco , k. grube , f. eilers , t. wolf , h. v. lhneysen , jps conf . proc . * 3 * , 015007 ( 2014 ) . t. terashima , m. kimata , n. kurita , h. satsukawa , a. harada , k. hazama , m. imai , a. sato , k. kihou , c .- h . lee , h. kito , h. eisaki , a. iyo , t. saito , h. fukazawa , y. kohori , h. harima , and s. uji , phys . jpn . * 79 * , 053702 ( 2010 ) . s. backes , d. guterding , h.o . jeschke , r. valenti , new j. phys . * 16 * , 083025 ( 2014 ) . our dft calculations did not show a lifshitz transition under pressure up to 4 gpa for all experimentally observed fermi surface sheets ( @xmath13 , @xmath86 , @xmath87 , and @xmath15 ) . we note that the small pillow at @xmath106 point observed at @xmath88 in the calculated band structure ( see fig . [ bands_pressure_pm ] ) disappears with the pressure . however , this particular fermi surface sheet was not observed experimentally @xcite , therefore in the analysis we ignored this pillow with a very small pdos .

the effect of hydrostatic pressure and partial na substitution on the normal - state properties and the superconducting transition temperature ( @xmath0 ) of k@xmath1na@xmath2fe@xmath3as@xmath3 single crystals were investigated . it was found that a partial na substitution leads to a deviation from the standard @xmath4 fermi - liquid behavior in the temperature dependence of the normal - state resistivity . it was demonstrated that non - fermi liquid like behavior of the resistivity for k@xmath1na@xmath5fe@xmath3as@xmath3 and some kfe@xmath3as@xmath3 samples can be explained by disorder effect in the multiband system with rather different quasiparticle effective masses . concerning the superconducting state our data support the presence of a shallow minimum around 2 gpa in the pressure dependence of @xmath0 for stoichiometric kfe@xmath3as@xmath3 . the analysis of @xmath0 in the k@xmath1na@xmath5fe@xmath3as@xmath3 at pressures below 1.5 gpa showed , that the reduction of @xmath0 with na substitution follows the abrikosov - gorkov law with the critical temperature @xmath6 of the clean system ( without pair - breaking ) which linearly depends on the pressure . our observations , also , suggest that @xmath0 of k@xmath1na@xmath2fe@xmath3as@xmath3 is nearly independent of the lattice compression produced by the na substitution . further , we theoretically analyzed the behavior of the band structure under pressure within the generalized gradient approximation ( gga ) . a qualitative agreement between the calculated and the recently in de haas - van alphen experiments @xcite measured pressure dependencies of the fermi - surface cross - sections has been found . these calculations , also , indicate that the observed minimum around 2 gpa in the pressure dependence of @xmath0 may occur without a change of the pairing symmetry .

news articles are very dynamic due to their relation to continuously developing events that typically have short lifespans . for a news article to be popular , it is essential for it to propagate to a large number of readers within a short time . hence there exists a competition among different sources to generate content which is relevant to a large subset of the population and becomes virally popular . traditionally , news reporting and broadcasting has been costly , which meant that large news agencies dominated the competition . but the ease and low cost of online content creation and sharing has recently changed the traditional rules of competition for public attention . news sources now concentrate a large portion of their attention on online mediums where they can disseminate their news effectively and to a large population . it is therefore common for almost all major news sources to have active accounts in social media services like twitter to take advantage of the enormous reach these services provide . due to the time - sensitive aspect and the intense competition for attention , accurately estimating the extent to which a news article will spread on the web is extremely valuable to journalists , content providers , advertisers , and news recommendation systems . this is also important for activists and politicians who are using the web increasingly more to influence public opinion . however , predicting online popularity of news articles is a challenging task . first , _ context _ outside the web is often not readily accessible and elements such as local and geographical conditions and various circumstances that affect the population make this prediction difficult . furthermore , _ network properties _ such as the structure of social networks that are propagating the news , influence variations among members , and interplay between different sections of the web add other layers of complexity to this problem . most significantly , intuition suggests that the _ content _ of an article must play a crucial role in its popularity . content that resonates with a majority of the readers such as a major world - wide event can be expected to garner wide attention while specific content relevant only to a few may not be as successful . given the complexity of the problem due to the above mentioned factors , a growing number of recent studies @xcite , @xcite , @xcite , @xcite , @xcite make use of early measurements of an item s popularity to predict its future success . in the present work we investigate a more difficult problem , which is prediction of social popularity without using early popularity measurements , by instead solely considering features of a news article _ prior _ to its publication . we focus this work on observable features in the content of an article as well as its source of publication . our goal is to discover if any predictors relevant only to the content exist and if it is possible to make a reasonable forecast of the spread of an article based on content features . the news data for our study was collected from feedzilla a news feed aggregator and measurements of the spread are performed on twitter , an immensely popular microblogging social network . social popularity for the news articles are measured as the number of times a news url is posted and shared on twitter . to generate features for the articles , we consider four different characteristics of a given article . namely : * the news source that generates and posts the article * the category of news this article falls under * the subjectivity of the language in the article * named entities mentioned in the article we quantify each of these characteristics by a score making use of different scoring functions . we then use these scores to generate predictions of the spread of the news articles using regression and classification methods . our experiments show that it is possible to estimate ranges of popularity with an overall accuracy of 84% considering only content features . additionally , by comparing with an independent rating of news sources , we demonstrate that there exists a sharp contrast between traditionally popular news sources and the top news propagators on the social web . in the next section we provide a survey of recent literature related to this work . section 3 describes the dataset characteristics and the process of feature score assignment . in section 4 we will present the results of prediction methods . finally , in section 5 we will conclude the paper and discuss future possibilities for this research . stochastic models of information diffusion as well as deterministic epidemic models have been studied extensively in an array of papers , reaffirming theories developed in sociology such as diffusion of innovations @xcite . among these are models of viral marketing @xcite , models of attention on the web @xcite , cascading behavior in propagation of information @xcite @xcite and models that describe heavy tails in human dynamics @xcite . while some studies incorporate factors for content _ fitness _ into their model @xcite , they only capture this in general terms and do not include detailed consideration of content features . @xcite performed a controlled experiment on music , comparing quality of songs versus the effects of social influence@xcite . they found that song quality did not play a role in popularity of highly rated songs and it was social influence that shaped the outcome . the effect of user influence on information diffusion motivates another set of investigations @xcite , @xcite,@xcite , @xcite . on the subject of news dissemination , @xcite and @xcite study temporal aspects of spread of news memes online , with @xcite empirically studying spread of news on the social networks of digg and twitter and @xcite studying facebook news feeds . a growing number of recent studies predict spread of information based on early measurements ( using early votes on digg , likes on facebook , click - throughs , and comments on forums and sites ) . @xcite found that eventual popularity of items posted on youtube and digg has a strong correlation with their early popularity ; @xcite and @xcite predict the popularity of a discussion thread using features based on early measurements of user comments . @xcite propose the notion of a virtual temperature of weblogs using early measurements . @xcite predict digg counts using stochastic models that combine design elements of the site -that in turn lead to collective user behavior- with information from early votes . finally , recent work on variation in the spread of content has been carried out by @xcite with a focus on categories of twitter hashtags ( similar to keywords ) . this work is aligned with ours in its attention to importance of content in variations among popularity , however they consider categories only , with news being one of the hashtag categories . @xcite conduct similar work on social marketing messages . this section describes the data , the feature space , and feature score assignment in detail . data was collected in two steps : first , a set of articles were collected via a news feed aggregator , then the number of times each article was linked to on twitter was found . in addition , for some of the feature scores , we used a 50-day history of posts on twitter . the latter will be explained in section [ sec : feature description and scoring ] on feature scoring . online news feed aggregators are services that collect and deliver news articles as they are published online . using the api for a news feed aggregator named feedzilla , we collected news feeds belonging to all news articles published online during one week ( august 8th to 16th , 2011 ) . the feed for an article includes a title , a short summary of the article , its url , and a time - stamp . in addition , each article is pre - tagged with a category either provided by the publisher or in some manner determined by feedzilla . a fair amount of cleaning was performed to remove redundancies , resolve naming variations , and eliminate spam through the use of automated methods as well as manual inspection . as a result over 2000 out of a total of 44,000 items in the data were discarded . the next phase of data collection was performed using topsy , a twitter search engine that searches all messages posted on twitter . we queried for the number of times each news link was posted or reshared on twitter ( tweeted or retweeted ) . earlier research @xcite on news meme buildup and decay suggest that popular news threads take about 4 days until their popularity starts to plateau . therefore , we allowed 4 days for each link to fully propagate before querying for the number of times it has been shared . the first half of the data was used in category score assignment ( explained in the next section ) . the rest we partitioned equally into 10,000 samples each for training and test data for the classification and regression algorithms . figure [ fig : powerlaw100 ] shows the log distribution of total tweets over all data , demonstrating a long tail shape which is in agreement with other findings on distribution of twitter information cascades @xcite . the graph also shows that articles with zero tweets lie outside of the general linear trend of the graph because they did not propagate on the twitter social network . our objective is to design features based on content to predict the number of tweets for a given article . in the next section we will describe these features and the methods used to assign values or scores to features . choice of features is motivated by the following questions : does the category of news affect its popularity within a social network ? do readers prefer factual statements or do they favor personal tone and emotionally charged language ? does it make a difference whether famous names are mentioned in the article ? does it make a difference who publishes a news article ? these questions motivate the choice of the following characteristics of an article as the feature space : the category that the news belongs to ( e.g. politics , sports , etc . ) , whether the language of the text is objective or subjective , whether ( and what ) named entities are mentioned , and what is the source that published the news . these four features are chosen based on their availability and relevance , and although it is possible to add any other available features in a similar manner , we believe the four features chosen in this paper to be the most relevant . we would like to point out that we use the terms article and link interchangeably since each article is represented by its url link . news feeds provided by feedzilla are pre - tagged with category labels describing the content . we adopted these category labels and designed a score for them which essentially represents a prior disribution on the popularity of categories . figure [ fig : categories ] shows a plot of categories and the number of article links in each category . we observe that news related to technology has a more prominent presence in our dataset and most probably on twitter as a whole . furthermore , we can see categories ( such as health ) with low number of published links but higher rates of tweet per link . these categories perhaps have a niche following and loyal readers who are intent on posting and retweeting its links . observing the variations in average tweets per link from figure [ fig : categories ] we use this quantity to represent the prior popularity for a category . in order to assign a value ( i.e. score ) to each category , we use the the first 22,000 points in the dataset to compute the average tweet per article link in that category . we call this average tweet per link the _ t - density _ score and we will use this measure in score assignments for some other features as well . another feature of an article that can affect the amount of online sharing is its language . we want to examine if an article written in a more emotional , more personal , and more subjective voice can resonate stronger with the readers . accordingly , we design a binary feature for subjectivity where we assign a zero or one value based on whether the news article or commentary is written in a more subjective voice , rather than using factual and objective language . we make use of a subjectivity classifier from lingpipe @xcite a natural language toolkit . since this requires training data , we use transcripts from well - known tv and radio shows belonging to rush limbaugh and keith olberman as the corpus for subjective language . on the other hand , transcripts from cspan as well as the parsed text of a number of articles from the website firstmonday are used as the training corpus for objective language . the above two training sets provide a very high training accuracy of 99% and manual inspection of final results confirmed that the classification was satisfactory . figure [ fig : subjectivity ] illustrates the distribution of average subjectivity per source , showing that some sources consistently publish news in a more objective language and a somwhat lower number in a more subjective language . in this paper , a named entity refers to a known place , person , or organization . intuition suggests that mentioning well - known entities can affect the spread of an article , increasing its chances of success . for instance , one might expect articles on obama to achieve a larger spread than those on a minor celebrity . and it has been well documented that fans are likely to share almost any content on celebrities like justin bieber , oprah winfrey or ashton kutcher . we made use of the stanford - ner entity extraction tool to extract all the named entities present in the title and summary of each article . we then assign scores to over 40,000 named entities by studying historical prominence of each entity on twitter over the timeframe of a month . the assigned score is the average t - density ( tweet per link ) of each named entity . to assign a score for a given article we use three different values : the number of named entities in an article , the highest score among all the named entities in an article , and the average score among the entities . the data includes articles from 1350 unique sources on the web . we assign scores to each source based on the historical success of each source on twitter . for this purpose , we collected the number of times articles from each source were shared on twitter in the past . we used two different scores , first the aggregate number of times articles from a source were shared , and second the t - density of each source ( average number of times each article belonging to a source was shared ) . the latter proved to be a better score assignment compared to the aggregate . to investigate whether it is better to use a smaller portion of more recent history , or a larger portion going back farther in time and possibly collecting outdated information , we start with the two most recent weeks prior to our data collection and increase the number of days , going back in time . figure [ fig : sourcecor ] shows the trend of correlation between the t - density of sources in historical data and their true t - density of our dataset . we observe that the correlation increases with more datapoints from the history until it begins to plateau near 50 days . using this result , we take 54 days of history prior to the first date in our dataset . we find that the correlation of the assigned score found in the above manner has a correlation of 0.7 with the t - density of the dataset . meanwhile , the correlation between the source score and number of tweets of any given article is 0.35 , suggesting that information about the source of publication alone is not sufficient in predicting popularity . figure [ fig : sourcescoredist ] shows the distribution of log of source scores ( t - density ) . taking the log of source scores produces a more normal shape , leading to improvements in regression algorithms . we plot the timeline of t - densities for a few sources and find that t - density of a source can vary greatly over time . figure [ fig : mashable ] shows the t - density values belonging to the technology blog _ mashable _ and _ blog maverick _ , a weblog of prominent entrepreneur , mark cuban . the t - density scores corresponding to each of these sources are 74 and 178 respectively . however , one can see that _ mashable _ has a more consistent t - density compared to _ blog maverick_. in order to improve the score to reflect consistency we devise two methods ; the first method is to smooth the measurements for each source by passing them through a low - pass filter . second is to weight the score by the percentage of times a source s t - density is above the mean t - density over all sources , penalizing sources that drop low too often . the mean value of t - densities over all sources is 6.4 . figure [ fig : temporal ] shows the temporal variations of tweets and links over all sources . notice that while both tweets and links have a weekly cycle , the t - density ( tweets over links ) does not have this periodic nature . + as we assign scores to sources in our dataset , we are interested to know whether sources successful in this dataset are those that are conventionally considered prominent . google news is one of the major aggregators and providers of news on the web . while inclusion in google news results is free , google uses its own criteria to rank the content and place some articles on its homepage , giving them more exposure . freshness , diversity , and rich textual content are listed as the factors used by google news to automatically rank each article as it is published . because google does not provide overall rankings for news sources , to get a rating of sources we use newsknife . newsknife is a service that rates top news sites and journalists based on analysis of article s positions on the google news homepage and sub - pages internationally . we would like to know whether the sources that are featured more often on google news ( and thus deemed more prominent by google and rated more highy by newsknife ) are also those that become most popular on our dataset . .correlation values between newsknife source scores and their performance on twitter dataset . [ cols="<,<,<,<",options="header " , ] in this work we predicted the popularity of news items on twitter using features extracted from the content of news articles . we have taken into account four features that cover the spectrum of the information that can be gleaned from the content - the source of the article , the category , subjectivity in the language and the named entities mentioned . our results show that while these features may not be sufficient to predict the exact number of tweets that an article will garner , they can be effective in providing a range of popularity for the article on twitter . we achieved an overall accuracy of 84% using classifiers . it is important to bear in mind that while it is intriguing to pay attention to the most popular articles those that become viral on the web a great number of articles spread in medium numbers . these medium levels can target highly interested and informed readers and thus the mid - ranges of popularity should not be dismissed . interestingly we have found that in terms of number of retweets , the top news sources on twitter are not necessarily the conventionally popular news agencies and various technology blogs such as mashable and the google blog are very widely shared in social media . overall , we discovered that one of the most important predictors of popularity was the source of the article . this is in agreement with the intuition that readers are likely to be influenced by the news source that disseminates the article . on the other hand , the category feature did not perform well . one reason for this is that we are relying on categories provided by feedzilla , many of which overlap in content . thus a future task is to extract categories independently and ensure little overlap . combining other layers of complexity described in the introduction opens up the possibility of better prediction . it would be interesting to incorporate network factors such as the influence of individual propagators to this work . agarwal , n. ; liu , h. ; tang , l. ; and yu , p. s. 2008 . identifying the influential bloggers in a community . in _ proceedings of the international conference on web search and web data mining _ , wsdm 08 , 207218 . new york , ny , usa : acm . kim , s .- d . ; kim , s .- h . ; and cho , h .- 2011 . predicting the virtual temperature of web - blog articles as a measurement tool for online popularity . in _ ieee 11th international conference on computer and information technology ( cit ) _ , 449 454 . romero , d. m. ; meeder , b. ; and kleinberg , j. 2011 . differences in the mechanics of information diffusion across topics : idioms , political hashtags , and complex contagion on twitter . in _ proceedings of the 20th international conference on world wide web _ , www 11 , 695704 . new york , ny , usa : acm . tatar , a. ; leguay , j. ; antoniadis , p. ; limbourg , a. ; de amorim , m. d. ; and fdida , s. 2011 . predicting the popularity of online articles based on user comments . in _ proceedings of the international conference on web intelligence , mining and semantics _ , wims 11 , 67:167:8 . new york , ny , usa : acm . zhou , z. ; bandari , r. ; kong , j. ; qian , h. ; and roychowdhury , v. 2010 . information resonance on twitter : watching iran . in _ proceedings of the first workshop on social media analytics _ , soma 10 , 123131 . new york , ny , usa : acm .

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ news articles are extremely time sensitive by nature . there is also intense competition among news items to propagate as widely as possible . hence , the task of predicting the popularity of news items on the social web is both interesting and challenging . prior research has dealt with predicting eventual online popularity based on early popularity . it is most desirable , however , to predict the popularity of items prior to their release , fostering the possibility of appropriate decision making to modify an article and the manner of its publication . in this paper , we construct a multi - dimensional feature space derived from properties of an article and evaluate the efficacy of these features to serve as predictors of online popularity . we examine both regression and classification algorithms and demonstrate that despite randomness in human behavior , it is possible to predict ranges of popularity on twitter with an overall 84% accuracy . our study also serves to illustrate the differences between traditionally prominent sources and those immensely popular on the social web . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

continuous - variable quantum key distribution ( cvqkd ) encodes information into the quadratures of optical fields and extracts it with homodyne detection , which has higher efficiency and repetition rate than that of the single photon detector @xcite . cvqkd , especially the gg02 protocol @xcite , is hopeful to realize high speed key generation between two parties , alice and bob . besides experimental demonstrations @xcite , the theoretical security of cvqkd has been established against collective attacks @xcite , which has been shown optimal in the asymptotical limit @xcite . the practical security of cvqkd has also been noticed in the recent years , and it has been shown that the source noise in state preparation may be undermine the secure key rate @xcite . in gg02 , the coherent states should be displaced in phase space following gaussian modulation with variance @xmath0 . however , due to the imperfections in laser source and modulators , the actual variance is changed to @xmath1 , where @xmath2 is the variance of source noise . an method to describe the trusted source noise is the beamsplitter model @xcite . this model has a good approximation for source noise , especially when the transmittance of beamsplitter @xmath3 approaches @xmath4 , which means that the loss in signal mode is negligible . however , this method has the difficulty of parameter estimation to the ancilla mode of the beamsplitter , without the information of which , the covariance matrix of the system are not able to determine . in this case , the optimality of gaussian attack @xcite should be reconsidered @xcite , and we have to assume that the channel is linear to calculate the secure key rate . to solve this problem , we proposed an improved source noise model with general unitary transformation @xcite . without extra assumption on quantum channel and ancilla state , we are able to derive a tight security bound for reverse reconciliation , as long as the variance of source noise @xmath2 can be properly estimated . the optimality of gaussian attack is kept within this model . the remaining problem is to estimate the variance of source noise properly . without such a source monitor , alice and bob can not discriminate source noise from channel excess noise , which is supposed to be controlled by the eavesdropper ( eve ) @xcite . in practice , source noise is trusted and is not controlled by eve . so , such _ untrusted source noise model _ just overestimates evepower and leads to an untight security bound . a compromised method is to measure the quadratures of alice s actual output states each time before starting experiment . however , this work is time consuming , and in qkd running time , the variance of source noise may fluctuate slowly and deviate from preliminary result . in this paper , we propose two real - time schemes , that the active switch scheme and the passive beamsplitter scheme to monitor the variance of source noise , with the help of which , we derive the security bounds asymptotically for both of them against collective attacks , and discuss their potential applications when finite size effect is taken into account . in this section , we introduce two real - time schemes to monitor the variance of source noise for the gg02 protocol , based on our general model . both schemes are implemented in the so - called prepare and measurement scheme ( p&m scheme ) @xcite , while for the ease of theoretical research , here we analyze their security in the entanglement - based scheme ( e - b scheme ) . the covariance matrix , used to simplify the calculation , is defined by @xcite @xmath5,\ ] ] where operator @xmath6 , @xmath7 , mean value @xmath8 $ ] , @xmath9 is the density matrix , and @xmath10 denotes the anticommutator . in e - b scheme , alice prepares epr pairs , measuring the quadratures of one mode with two balanced homodyne detectors , and then send the other mode to bob . it is easy to verify that the covariance matrix of an epr pair is @xmath11 where @xmath12 is the variance of the epr modes , and @xmath13 corresponds to alice s modulation variance in the p&m scheme . however , due to the effect of source noise , the actual covariance matrix is changed to @xmath14 where @xmath2 is the variance of source noise . as mentioned in @xcite , we assume this noise is introduced by a neutral party , fred , who purifies @xmath15 and introduces the source noise with arbitrary unitary transformation . in this section , we show how to monitor @xmath2 with our active and passive schemes , and derive the security bounds in the infinite key limit . a method of source monitoring is to use an active optical switch , controlled by a true random number generator ( trng ) , combined with a homodyne detection . the entanglement - based version @xcite of this scheme is illustrated in fig . [ pic2 ] , where we randomly select parts of signal pulses , measure their quadratures and estimate their variance . in the infinite key limit , the pulses used for source monitor should have the same statistical identities with that sent to bob . comparing the estimated value with the theoretical one , we are able to derive the variance of source noise , and the security bound can be calculated by @xcite @xmath16,\ ] ] where @xmath17 is the sampling ratio of source monitoring , @xmath18 is the classical mutual information between alice and bob , @xmath19 is the quantum mutual information between eve and bob , and @xmath20 is the reconciliation efficiency . after channel transmission , the whole system can be described by covariance matrix @xmath21 @xmath22 where @xmath2 is the variance of source noise , @xmath23 matrix @xmath24 is related to fred s two - mode state , @xmath25 is the transmittance and @xmath26 is the channel noise and @xmath27 is the channel excess noise . in practice , the covariance matrix can be estimated with experimental data with source monitor and parameter estimation . here , for the ease of calculation , we assume that parameters @xmath25 and @xmath27 have known values . in the infinite key limit . , height=144 ] given @xmath21 , the classical mutual information @xmath18 can be directly derived , while @xmath19 can not , since the ancilla state @xmath28 is unknown in our general model . fortunately , we can substitute @xmath21 with another state @xmath29 when calculating @xmath19 @xcite , where @xmath30}\sigma_{z } \\ 0 & 0 & \sqrt{\eta [ ( v+\chi_{s})^{2}-1]}\sigma_{z } & \eta ( v+\chi_{s}+\chi)\mathbb{i } \\ \end{array } \right),\ ] ] and we have shown that such substitution provides a tight bound for reverse reconciliation . here , we have assumed that the pulses generated in alice is i.i.d . , and the true random number plays an important role in this scheme , without which , the sampled pulses may have different statistical characters from signal pulses sent to bob . the asymptotic performance of this scheme will be analyzed in sec . though the active switch scheme is intuitive in theoretical research , it is very not convenient in the experimental realization , since the high speed optical switch and an extra trng are needed . also , it lowers the secure key rate with @xmath31 due to the sampling ratio . inspired by @xcite , we propose a passive beam splitter scheme to simplify the implementation . as illustrated in fig . [ bsscheme ] , a beamsplitter is used to separate mode @xmath32 into two parts . one mode , @xmath33 , is monitored by alice , and the other , @xmath34 , is sent to bob . is replaced by a beamsplitter . alice and bob are able to estimate the source noise by measuring mode @xmath33 with the homodyne detection . , height=144 ] the security bound of passive beam splitter scheme can be calculated in a similar way that we substitute the whole state @xmath35 with @xmath36 . the covariance of its subsystem , @xmath37 , can be written as @xmath38 where mode @xmath39 is initially in the vacuum state . the covariance matrix after beam splitter is @xmath40 where @xmath41 then , mode @xmath34 is sent to bob through quantum channel , characterized by @xmath42 . the calculation of @xmath19 in this scheme is a little more complex , since an extra mode @xmath33 is introduced by the beamsplitter . we omit the detail of calculation here , which can be derived from @xcite . the performance of this scheme is discussed in the next section . in this section , we analyze the performance of both schemes with numerical simulation . as mentioned above , the simulation is restricted to the asymptotic limit . the case of finite size will be discussed later . to show the performance of source monitor schemes , we illustrate the secure key rate in fig . [ comparision ] , in which the _ untrusted noise scheme _ is included for comparison . for the ease of discussion , the imperfections in practical detectors are not included in our simulation , the effect of which have been studied previously @xcite . , the source noise is @xmath43 , the channel excess noise is @xmath44 , and the reconciliation efficiency is @xmath45 . the sample ration in the optical switch scheme is @xmath46 , and the transmittance in the beam splitter scheme is @xmath47.,height=288 ] as shown in fig . [ comparision ] , secure key rate of each scheme is limited within @xmath48 , where large excess noise @xmath49 is used . under state - of - the - art technology , the excess noise can be controlled less than a few percent of the shot noise . so , our simulation is just a conservative estimation on the secure key rate . the _ untrusted source noise scheme _ has the shortest secure distance , because it ascribes the source noise into channel noise , which is supposed to be induced by the eavesdropper . in fact , source noise is neutral and can be controlled neither by alice and bob , nor by eve . so , this scheme just overestimates eve s power by supposing she can acquire extra information from source noise , which lower the secure key rate of this scheme . both the active and passive schemes have longer secure distance than the untrusted noise scheme , since they are based on the general source noise model , which does not ascribe source noise into eve s knowledge . the active switch scheme has lower secure key rate in the short distance area . this is mainly because that the random sampling process intercepts parts of the signal pulses to estimate the variance of source noise , which reduces the repetition rate with ratio @xmath17 . nevertheless , it does not overestimate eve s power @xcite . as a result , the secure key distance is improved . and transmittance @xmath50 , in which t varies from @xmath51 to @xmath52 . the colored parts illustrates the area with positive secure key rate , the empty parts illustrates to insecure area , and the abscissa values of boundary points corresponds to the secure distance.,height=336 ] both the secure key rate and secure distance of beam splitter scheme are superior than that of other schemes , when the transmittance is set to be @xmath53 , equal to the sampling rate @xmath17 in optical switch scheme , where no extra vacuum noise is introduced . this phenomena is quite similar to the `` noise beat noise '' scheme @xcite , which improves the secure key rate by introduce an extra noise into bob s side . though such noise lowers the mutual information between alice and bob , it also makes eve more difficult to estimate bob s measurement result . with the help of simulation , we find a similar phenomenon in the beam splitter scheme , the vacuum noise reduces mutual information @xmath54 more rapidly than its effect on @xmath18 . a preliminary explanation is that the sampled pulse in optical switch scheme is just used to estimate the noise variance , while in beam splitter model it increases eve s uncertainty on bob s information . combined with advantages in experimental realization , beam splitter scheme should be a superior choice . to optimize the performance of passive beam splitter scheme , we illustrate the secure key rate in fig . [ 3d ] for different beam splitter transmittance @xmath3 . the maximal secure distance about @xmath55 is achieved when @xmath56 , about @xmath57 km longer than that when @xmath58 . combined with the discussion above , this result can be understood as a balance between the effects of the noise on @xmath18 and @xmath54 , induced by beamsplitter . when @xmath3 is too small , @xmath18 also decreases rapidly , which limits the secure distance . the performance of source monitor schemes above is analyzed in asymptotical limit . in practice , the real - time monitor will concern the finite - size effect , since the variance of source noise may change slowly . a thorough research in finite size effect is beyond the scope of this paper , because the security of cvqkd in finite size is still under development , that the optimality of gaussian attack and collective attack has not been shown in the finite size case . nevertheless , we are able to give a rough estimation on the effect of block size , for a given distance . taking the active optical switch scheme for example , with a similar method in @xcite , the maximum - likelihood estimator @xmath59 is given by @xmath60 where @xmath61 , @xmath62 is the measurement result of source monitor , and @xmath63 is the expected value of the variance of source noise . for large @xmath64 , the @xmath65 distribution converges to a normal distribution . so , we have @xmath66 where @xmath67 is such that @xmath68 , and @xmath69 is the failure probability . the reason why we choose @xmath70 is that given the values of @xmath71 and @xmath25 , estimated by bob , the minimum of @xmath2 corresponds to the maximum of channel noise @xmath72 , which may be fully controlled by eve . the extra variance @xmath73 due to the finite size effect in source monitor is @xmath74 for @xmath75 , we have @xmath76 . as analyzed in @xcite , if the distance between alice and bob is @xmath77 km ( @xmath78 ) , the block length should be at least @xmath79 , which corresponds to @xmath80 induced by the finite size effect in source monitor . compared with the channel excess noise of @xmath81 , the effect of finite size in source monitor is very slight . due to the high repetition rate in cvqkd , alice and bob are able to accumulate such a block within several minutes , during which the source noise may change slightly . in conclusion , we propose two schemes , the active optical switch scheme and the passive beamsplitter scheme , to monitor the variance of source noise @xmath2 . combined with previous general noise model , we derive tight security bounds for both schemes with reverse reconciliation in the asymptotic limit . both schemes can be implemented under current technology , and the simulation result shows an better performance of our schemes , compared with the untrusted source noise model . further improvement in secure distance can be achieved , when the transmittance @xmath3 is optimized . in practise , the source noise varies slowly . to realize real - time monitoring , the finite size effect should be taken into account , that the block size should not be so large , that the source noise has changed significantly within this block , and the block size should not be too small , that we can not estimate the source noise accurately . the security proof of cvqkd with finite block size has not been established completely , since the optimality of collective attack and gaussian attack has not been shown in finite size . nevertheless , we derive the effective source noise induced by the finite block size , and find its effect is not significant in our scheme . so , our schemes may be helpful to realize real - time source monitor in the future . this work is supported by the key project of national natural science foundation of china ( grant no . 60837004 ) , national hi - tech research and development ( 863 ) program . the authors thank yujie shen , bingjie xu and junhui li for fruitful discussion . 10 scarani v , bechmann - pasquinucci h , cerf n j , duek m , ltkenhaus n , and peev m 2009 _ rev . mod . phys . _ * 81 * 1301

the noise in optical source needs to be characterized for the security of continuous - variable quantum key distribution ( cvqkd ) . two feasible schemes , based on either active optical switch or passive beamsplitter are proposed to monitor the variance of source noise , through which , eve s knowledge can be properly estimated . we derive the security bounds for both schemes against collective attacks in the asymptotic case , and find that the passive scheme performs better .

weyl fermions are chiral massless fermions @xcite . they can appear as quasiparticles in condensed - matter systems with strong spin - orbit coupling , where accidental band crossings cause a dirac - like dispersion around discrete points in the brillouin zone @xcite . in combination with a broken inversion or time - reversal symmetry , dirac points split into pairs of weyl points with opposite chirality . the term chirality denotes the fact that the weyl points are either sources or drains of berry flux . in the presence of parallel magnetic and electric fields an imbalance in the number of particles of each chirality @xcite leads to the so - called chiral magnetic effect resulting in a negative longitudinal magnetoresistance ( @xmath0 ) @xcite . recently , weyl fermions were predicted near the fermi energy in the non - centrosymmetric semimetals taas , tap , nbas , and nbp @xcite . band - structure calculations were confirmed by quantum oscillation studies and angle - resolved photoemission spectroscopy , where linear band crossings and fermi - arc surface states were discovered @xcite . furthermore , a negative longitudinal @xmath1 was detected for taas @xcite , nbas @xcite , tap @xcite , nbp @xcite and interpreted as the chiral magnetic effect . similar results have been reported for the dirac semimetals na@xmath2bi @xcite , zrte@xmath3 @xcite , and cd@xmath4as@xmath2 @xcite . however , a negative longitudinal @xmath1 was also observed in materials , which where not expected to show a chiral magnetic effect , _ e.g. _ in mpn@xmath4 ( m = nb , ta ; pn = as , sb ) @xcite , in gdptbi @xcite , or in the highly conductive delafossites pdcoo@xmath4 and ptcoo@xmath4 and in the ruthenate sr@xmath4ruo@xmath5 @xcite . all these materials show a high charge carrier mobility which causes a magnetic field - induced resistance anisotropy @xmath6 . due to the orbital effect in high - mobility materials , the resistivity normal to the magnetic field @xmath7 increases significantly upon increasing field , whereas the resistivity parallel to the magnetic field @xmath8 is not affected . other materials with a strong orbital @xmath1 are , for example , pure elemental bismuth , tungsten , or chromium @xcite . in these materials , it is well known that an inhomogeneous current distribution can occur in magnetic fields , when the current is not injected homogeneously into the sample @xcite . for point - like current contacts and a large resistance anisotropy @xmath9 ( induced by the magnetic field ) , the current flows predominantly parallel to the magnetic field direction , where the resistance is small . if the magnetic field @xmath10 is applied along the line connecting the current contacts , the electrical current forms a jet between them , hence the name `` current jetting '' @xcite . in that case , the voltage probes on the surface of the sample decouple from the current and probe a smaller or even zero potential difference @xmath11 @xcite . as the resistance anisotropy grows upon increasing field , the measured voltage decreases leading to an apparent negative longitudinal @xmath1 , defined here by @xmath12 , @xmath13 . recently , indications of an inhomogeneous current distribution were reported in tap @xcite . in this paper , we provide evidence for a field - induced current - jetting effect in the weyl semimetal candidates nbp , nbas , tap , and taas . we show that this effect can easily dominate the measurements of the longitudinal @xmath1 . the apparent negative longitudinal @xmath1 induced by the current jetting is so strong that it might hide a potentially existing chiral magnetic effect in this class of materials . the inhomogeneous current distribution associated with it manifests itself in additional characteristics , such as a strong dependence of the @xmath1 on the contact geometry @xcite as well as a typical angular dependence of the @xmath1 as a function of the field direction @xcite . even a``negative resistance '' might be observed , if the sample is not _ perfectly _ aligned with respect to the magnetic field @xcite . all experimental findings are in good agreement with finite - element simulations of the potential distribution taking into account the actual sample and contact geometries as well as a field - induced resistance anisotropy originating only from the transverse orbital @xmath1 . our study indicates that measurements of the longitudinal @xmath1 in high mobility materials with field - induced resistance anisotropy have to be verified carefully in order to rule out any influence of an inhomogeneous current distribution on the data . high - quality single crystals of nbp , nbas , tap , and taas were synthesized via a chemical vapor transport reaction . more details on the sample preparation and characterization can be found elsewhere @xcite . for the electrical resistance measurements , bar - shaped samples were cut from large oriented single crystals . typical dimensions of the samples were @xmath14 mm ( width @xmath15 thickness @xmath15 length ) . the electrical current ( @xmath16 ma ) was applied along the long sample direction and the resistance was measured in a 4probe geometry . we note that we define the current direction by the direction of the straight line connecting the two current electrodes . the contacts were made by spot welding platinum wires with a diameter of @xmath17 to the sample . the magneto - transport experiments were performed at @xmath18 k in magnetic fields up to @xmath19 t utilizing a physical property measurement system ( quantum design inc . ) . the resistance data were collected in positive and negative fields and subsequently symmetrized in order to remove a hall contribution present due to a small misalignment of the voltage contacts . the residual resistivity of the samples used in this work was about @xmath20 cm ( nbp ) , @xmath21 cm ( nbas ) , @xmath22 cm ( tap ) and @xmath23 cm ( taas ) . simulations of the potential distribution inside the samples were performed within a model based on current conservation using a finite element method , as implemented in the commercial software package comsol multiphysics 5.2 . for the simulations , we used a constant electrical current of @xmath16 ma , which was injected in the front and end surfaces ( @xmath24 planes of the tetragonal crystal structure ) near the top of the sample by point - like contacts . among the weyl semimetal candidate materials of the taas family , nbp is the one with the highest reported transverse magnetoresistance , 8500 at 1.85 k and 9 t , and an extremely large mobility of @xmath25 @xcite . this makes nbp best suited to study effects of the resistance anisotropy on the current distribution . in addition , in nbp the charge carriers at the fermi surface are only comprised of topologically trivial electrons as its weyl nodes lie 5 mev above the fermi energy and chirality is ill - defined @xcite . hence , a negative longitudinal @xmath26 , can not be ascribed to the chiral magnetic effect @xcite . in the following we will discuss the observation of a negative longitudinal @xmath1 in nbp , nbas , tap , and taas . first , we will describe its main origin based on an inhomogeneous current distribution , as caused by a magnetic field - induced resistance anisotropy , on the example of nbp , followed by a discussion of nbas , tap , and taas . finally , we will address further implications for the search of the chiral magnetic effect in weyl semimetals . , @xmath27 , and @xmath28 axes of the tetragonal crystal structure are in parallel to the sample edges as indicated . ( b ) magnetoresistance of nbp at 2 k for different angles of the magnetic field @xmath10 , as indicated . @xmath29 is the measured voltage for the contact pair @xmath30 normalized by the zero field value @xmath31 . the field is rotated in the @xmath24 plane . ] to investigate the effect of the magnetic field on the current distribution in nbp , we set up a device where the electrical current is injected at the corner of the sample and two pairs of voltage contacts @xmath30 and @xmath32 are attached at different positions relative to the current electrodes . the geometry of our nbp sample with point - like electrodes is shown in fig . [ mr]a . in zero field , the values of @xmath30 and @xmath32 are identical within the experimental uncertainty when normalized by the contact separation . the magnetic field dependence of the mr is depicted in fig . [ mr]b for different angles @xmath33 . here , @xmath33 is defined as the angle between the magnetic field and the long sample axis ( @xmath34 ) in the @xmath24 plane . while we observe a different longitudinal @xmath35 for the two voltage contacts pairs @xmath30 and @xmath32 , which we will discuss in details below , the transverse @xmath36 are identical . the transverse @xmath1 increases by a factor of 2500 up to 9 t at 1.85k and exhibits strong quantum oscillations . this demonstrates the high crystal quality and reflects the high mobility of the charge carriers . we note that our findings are in good agreement with the previously reported data @xcite . the mobility estimated from the quadratic field dependence of the resistance at low fields is @xmath37 . this is slightly smaller than the reported values in reference @xcite in agreement with the slightly lower @xmath1 . ) . the fact that the two contact pairs exhibit a different field dependence reflects an inhomogeneous current distribution in the sample . ] in the longitudinal configuration , where the magnetic field is applied parallel to the direction of the electrical current , @xmath38 and @xmath39 have a different magnetic - field dependence ( see fig . [ longitudinal_mr ] ) . this observation is a direct hint at an inhomogeneous current distribution inside the sample induced by the magnetic field . upon increasing the field @xmath40 corresponding to the contact pair far away from the current electrodes first increases slightly , before the voltage drops to a very small value above 1 t . on the other hand , for the voltage pair @xmath32 , which is almost in line with the current contacts , the voltage initially increases much stronger than @xmath40 until it reaches a maximum at around 0.5 t , indicating a high current flow in this region of the sample at small fields . upon further increasing @xmath10 , @xmath41 decreases rapidly displaying a minimum at around 3 t before a weak broad maximum appears . we note that for the same field @xmath41 is always considerably larger than @xmath40 ( except for zero magnetic field ) . a similar field dependence of the longitudinal voltage , particularly the strong decrease upon increasing magnetic field , has been reported in chromium @xcite , tungsten @xcite , and bismuth @xcite , and was explained by the current jetting effect @xcite . in order to get deeper insights into the influence of the magnetic field on the current distribution in nbp , we performed finite - element simulations of the potential distribution as a function of the field - induced resistance anisotropy . , 10 , 100 , and 1000 for a sample with point - like current electrodes and the dimensions of the investigated nbp sample . note , the current contacts have been moved from the corner to the middle compared with the experimental set up for a better visualization . here @xmath10 is aligned with the current electrodes , @xmath8 ( @xmath7 ) is the resistance parallel ( perpendicular ) to @xmath10 . the lines are contour lines of the equipotential surfaces . the increase in @xmath7 and hence the anisotropy , induced by an increase in the magnetic field , strongly distorts the equipotential lines . ] the potential distribution for various resistance anisotropies @xmath9 is depicted in fig . [ simulation ] . the sample has the same dimensions as the one studied experimentally , but the point - like current contacts have been attached in the middle of the front and end surfaces , close to the top of the sample for a better illustration . the magnetic field producing the resistance anisotropy is assumed to be parallel to the line connecting the current electrodes . @xmath42 corresponds to the initial potential distribution for @xmath43 in zero magnetic field . the equipotential planes , except for a region close to the current contacts , are parallel indicating a homogeneous current distribution in most parts of the sample . upon increasing the resistance anisotropy @xmath9 by increasing @xmath7 and keeping @xmath8 constant , we observe a dramatic effect on the potential distribution . even for a small anisotropy of @xmath44 , we find a pronounced distortion of the equipotential planes inside the sample leading to a strongly reduced region of a quasi - homogeneous potential distribution . for higher anisotropies , e.g. @xmath45 and @xmath46 , the equipotential planes are highly distorted as predicted for the current jetting effect @xcite . in order to extract the magnetic field dependence of the voltage from the simulations , we assume a field independent longitudinal resistance @xmath47 . in that case the anisotropy is defined as @xmath48 as given in fig . [ mr ] . note that for a magnetic field along the crystallographic @xmath34 axis , the transverse components of the resistance are generally not the same due to an anisotropy in the transverse mr in the crystallographic @xmath49 plane . for our simulations , we neglect this anisotropy and assume @xmath50 . the simulations to determine @xmath51 were carried out for the geometry of our sample and the actual positions of the contacts , in particular , the current was injected in the corners of the front and end surfaces in contrast to the visualization in fig . [ simulation ] . hence , @xmath52 corresponds to @xmath53 , @xmath54 to 0.2 t , @xmath55 to 0.5 t and @xmath56 to 3.2 t . the calculated field dependence of the voltage for the two different configurations is shown as dashed lines in fig . [ longitudinal_mr ] . it reflects purely the effect of a field induced transverse resistance on the longitudinal measurements . the agreement between the simulations and the experiment is astonishing . based on the results of our simulations , we argue that the pronounced voltage drop taking place upon increasing the magnetic field is caused by the intense increase of the transverse @xmath1 ( and correspondingly of the anisotropy @xmath9 ) , which forces the current to concentrate along a straight line connecting the current electrodes . voltage electrodes on the edge of the sample far away from this line , such as @xmath30 , detect a reduced voltage already for very small magnetic fields ( small anisotropies ) . in our geometry even an anisotropy of only @xmath54 ( @xmath57 t ) is enough to reduce the voltage to 65 % of the zero - field value . at the same time , the effective cross section of the sample is reduced , leading to higher currents and voltages close to the current jet . this is reflected in the pronounced increase of @xmath32 below 1 t. for higher fields ( higher anisotropy @xmath9 ) , the dilution even reaches the position of @xmath32 . we point out that our simulations , based on a simple current conservation model , reproduce quantitatively all the main experimental features , namely the hump at low magnetic fields and the apparent negative longitudinal @xmath26 . this suggests that the current jetting effect is the dominant cause of the experimentally observed _ negative _ longitudinal @xmath26 . between the magnetic field and the electrical current for the two contact pairs @xmath30 and @xmath32 , respectively . the inset in ( b ) displays the region where a negative voltage is observed for the contact pair @xmath32 . ] in the following , we will focus on the effect of a misalignment between the magnetic field and the current directions . in our experimental setup , the magnetic field is tilted away from the longitudinal configuration ( @xmath58 , @xmath59 ) towards the transverse configuration ( @xmath60 , @xmath61 ) @xmath62 is defined as in fig . special care was given to the alignment in the perpendicular direction , which is fixed on the rotator . figure [ neg_volt ] illustrates the effect of a small misalignment on @xmath38 and @xmath39 for angles up to @xmath63 . we point out the existence of strong characteristic differences in the field dependence of @xmath38 and @xmath39 in the small angle region . for @xmath38 a very small angular misalignment of @xmath64 is enough to obtain an increase in the voltage upon increasing the magnetic field ( see fig . [ neg_volt]a ) . this is due to the projection of the transverse mr to the current direction . moreover , for angles above @xmath65 the initial voltage drop present in the longitudinal configuration ( @xmath66 ) starts to disappear . in the case of @xmath32 a misalignment leads to a characteristically different field dependence . most strikingly , negative voltages are probed for misalignments between 0.5@xmath67 and 3@xmath67 in magnetic fields above 1 t . a similar observation has been reported recently for the longitudinal @xmath26 of tap , a member of the same family of materials @xcite . we will discuss the physical origin of the negative voltage values below . for nbp in @xmath19 t. ( b ) simulated potential distribution for @xmath68 ( applied field of @xmath69 t , @xmath70 , in nbp ) for different angles @xmath33 in the @xmath24 plane . this corresponds to a side view of the sample in fig . the current contacts have been attached on the front and end surfaces close to the middle of the edge parallel to @xmath28 for a better illustration . ( c ) schematic illustration of the relation between the angles , where a dip ( @xmath71 , @xmath72 ) and a hump ( @xmath73 , @xmath74 ) appear in @xmath75 and the geometry of the sample and positions of the electrodes . ] the angular dependence of @xmath32 in a constant magnetic field of 9 t is displayed in fig . [ angular_dependence]a . a negative voltage is detected , when the field direction is slightly tilted away from the longitudinal configuration . a dip appears in @xmath76 at an angle @xmath71 ( @xmath72 ) . upon further increasing @xmath33 , @xmath76 increases rapidly and a hump builds up at an angle @xmath73 ( @xmath74 ) . a similar angular dependence has been reported and explained for bi and sb samples in a similar geometry @xcite . the strong variations in @xmath75 near the transverse configuration ( @xmath77 ) stem from quantum oscillations @xcite . as shown in fig . [ simulation ] , the equipotential lines are strongly distorted in the presence of a high magnetic field . the potential distribution from our simulation for different angles between the magnetic field and the current direction is shown in fig . [ angular_dependence]b for @xmath68 . as described before , the magnetic field is tilted in the @xmath24 plane . figures [ angular_dependence]b and c show a side view of the sample . for a better illustration , the current is injected in the middle of the side surfaces parallel to the @xmath24 plane . once the magnetic field is tilted away from the line connecting the current contacts , the equipotential lines start to tilt and to align parallel to the direction of the magnetic field @xcite . for high magnetic fields , corresponding to a large resistance anisotropy @xmath9 , @xmath78 and @xmath79 are only determined by the geometry of the sample and the position of the electrodes as depicted in fig . [ angular_dependence]c @xcite . when the magnetic field is oriented parallel to the line connecting the positive current electrode @xmath80 and the closest voltage electrode @xmath81 , this contact will be at a maximum of the potential and we observe a hump in the angular dependence ( @xmath73 in fig . [ angular_dependence]a ) . similarly , when the magnetic field points towards the second voltage contact , the measured potential difference becomes negative and we observe a dip in the angular dependence ( @xmath71 in fig . [ angular_dependence]a ) . in our experiment , the agreement between the angles expected from the geometry , giving @xmath82)@xmath83 and @xmath84 ( marked by arrows in fig . [ angular_dependence]a ) and the position of the dips and humps in @xmath75 is very good . the appearance and the height of the dip and the hump also depend strongly on the size of @xmath9 , the thickness and the length of the sample , and the position of the voltage contacts with respect to the tilting plane of the magnetic field crossing the current electrodes @xcite . the dip and hump are most pronounced for voltage contacts lying in that plane ( such as @xmath32 ) and disappear for voltage contacts far away ( such as @xmath30 , not shown here ) @xcite . the dips also vanish for very thin samples and become shallow for small @xmath9 @xcite . in conclusion , dips and humps can appear in the angular dependence of @xmath85 characteristic of current jetting . for large resistance anisotropies , their position is only determined by the geometry of the sample and the position of the electrodes . this is a further evidence that the increase of the transverse resistance upon increasing field strongly influences any @xmath26 measurements in nbp and in similar materials . in general , for the longitudinal mr , the requirements on the alignment between the magnetic field and the current directions are extremely stringent ( misalignment < 0.25 @xmath67 in our setup ) . otherwise , negative voltages can be observed . the previous results , obtained on the example of nbp , are principally valid for all materials with a high mobility of the charge carriers and a field - induced anisotropy in the @xmath26 . this in particular applies for the closely related materials of the same family : tap , taas , and nbas . in ( a ) nbp , ( b ) tap , ( c ) taas and ( d ) nbas in the longitudinal configuration ( @xmath86 ) . two voltage contact pairs @xmath30 ( red lines ) and @xmath32 ( blue lines ) at different positions on the sample show a different field dependence due to an inhomogeneous current distribution . in the first case ( solid lines ) point - like current electrodes were spot - welded to the samples and in the second case ( dashed lines ) the current electrodes were attached with silver paste covering the whole front and end surfaces of the samples . ] figure [ fig6 ] displays the `` longitudinal @xmath87 '' for the four members of the taas family measured with two different voltage contact pairs and two differently prepared pairs of current electrodes . signatures of an inhomogeneous current distribution are present in all samples . the experimental setups with the spot - welded current contacts ( solid lines ) follow the schematic configuration illustrated in fig . the corresponding @xmath88 curves for the different voltage contact configurations @xmath30 and @xmath32 show a qualitative different field dependence . here , @xmath30 is situated far from the line connecting the current electrodes and @xmath32 close to it . none of these curves reflects the intrinsic longitudinal @xmath26 . note that , in the case of taas , the increase of @xmath38 and @xmath39 in high fields points at a small misalignment between the current and the magnetic field directions ( see fig . [ fig6]c ) . however , the question remains , how the intrinsic longitudinal @xmath26 can be probed in order to extract the information if the chiral magnetic effect is present in these materials . there are two obvious experimental approaches improving the homogeneity of the current distribution , and a combination of both might give the best results . the first is to inject the current homogeneously over the whole cross section of the sample . it turnes out that this is non - trivial . soldering contacts is not possible in these materials and neither silver paste nor silver epoxy resulted in a satisfying outcome . the dashed lines in fig . [ fig6 ] mark the longitudinal voltages obtained with same voltage pairs as before , but using current electrodes made by silver paste applied over the whole front and end surfaces of the samples . there is still a significant difference between @xmath38 and @xmath39 for all materials . this is caused by a local variation of the contact resistance of the silver paint contacts leading to an injection of the current into the sample where the contact resistance is smallest . since these positions are not necessarily the same on the front and end of the sample , the current direction is likely to be misaligned with respect to the direction of the magnetic field applied parallel to the sample . this might have been the case for experiments on nbp and taas ( fig . [ fig6]a and c ) where a negative voltage was detected ( see also sec . [ angdep ] ) . the second approach is to make the sample longer and/or thinner , so that the current has enough length to spread homogenously into the whole cross section of the sample . the critical parameter here is the aspect ratio @xmath89 of the sample length @xmath90 and width @xmath91 ( @xmath92 ) . it is well known that the resistance anisotropy can be viewed as an effective shortening of the sample in the direction of the magnetic field by a factor of @xmath93 ( or an increase of @xmath91 by a factor of @xmath94 ) @xcite . given a certain anisotropy of the resistivity @xmath9 , a suitable ratio of @xmath89 for a homogeneous current distribution without resistance anisotropy ( @xmath42 ) has to be be multiplied by a factor of @xmath94 in order to still ensure a homogeneous current distribution . this is difficult to achieve for high @xmath9 , which is determined in first approximation by the transverse mr . for example , in the investigated nbp sample with @xmath68 , the required @xmath89 ratio would be higher than 50 . additionally , our simulations show that these long thin samples are much more sensitive to a misalignment of the magnetic field and electrical current directions . on the other hand , our experimental data and simulations show that an inhomogeneous current distribution appears already at very low fields ( very small anisotropy @xmath9 ) . hence , for small ratios of @xmath89 ( @xmath95 for our sample of nbp ) even a resistance anisotropy below 10 can be large enough to reduce significantly the region in the sample where the current is evenly distributed . as a result , an inhomogeneous current distribution can appear already at high temperatures , very low magnetic fields , and also in materials with low mobilities . we have studied the magnetoresistance of the putative weyl semimetals of the taas family , where the chiral anomaly is expected to induce a negative longitudinal @xmath26 . indeed , an apparent negative @xmath26 is present in all investigated samples , nbp , nbas , tap , and taas . however , the field dependence of the _ measured _ longitudinal @xmath26 is highly sensitive to the sample geometry and the position of the electrical contacts , in particular of the voltage contacts . furthermore , it is extremely susceptible to the alignment between magnetic field and current directions . small deviations from a parallel alignment even lead to an apparent _ negative _ resistance . our experimental findings provide strong evidence for an inhomogeneous current distribution as the origin of the observed phenomena and seem to exclude the chiral magnetic effect as predominant cause . this inhomogeneous current distribution emanates from the field - induced resistance anisotropy in these high mobility materials and is known as current - jetting effect . our conclusion is further supported by finite - element simulations based on a simple current - conservation model with the resistance anisotropy as only free parameter , which agree exceptionally well with the experimental results . an important implication of the simulations is that the intrinsic longitudinal magnetoresistance is almost constant in changing magnetic fields . we point out that the inhomogeneous current distribution already appears for very small resistance anisotropies below 10 , which corresponds to small applied magnetic fields in our experiments . moreover , our simulations expose that for high anisotropies it is very difficult to avoid signatures of the current - jetting effect even in samples with optimized geometry and ideal electrical contacts . to conclude , our study demonstrates that measurement of the longitudinal @xmath26 in materials with field - induced resistance anisotropy is not straightforward and careful checking is required before intrinsic physical properties of the material , such as the chiral anomaly in weyl semimetals , are extracted . we acknowledge stimulating discussions with j. bardarson , k. behnia , m. brando , c. felser , a. grushin , b. yan and especially k. yoshida . we thank h. borrmann for orienting the single crystals . r. dos reis acknowledges financial support from the brazilian agency cnpq ( brazil ) . 99 arnold f , shekhar c , wu s - c , sun y , dos reis r d , kumar n , naumann m , ajeesh m o , schmidt m , grushin a g , bardarson j h , baenitz m , sokolov d , borrmann h , nicklas m , felser c , hassinger e , and b. yan 2016 _ nat . comm . _ * 7 * 11615 xu s y , belopolski i , alidoust n , neupane m , bian g , zhang c , sankar r , chang g , zhujun y , lee c c , shin - ming h , zheng h , ma j , sanchez d s , wang b , bansil a , chou f , shibayev p p , lin h , jia s , and hasan m z 2015 _ science _ * 349 * 613 xu s y , belopolski i , sanchez d s , zhang c l , chang q q , guo c , bian g , yuan z j , lu h , chang t r , shibayev p p , prokopovych m l , alidoust n , zheng h , lee c c , huang s m , sankar r , chou f c , hsu c h , jeng h t , bansil a , neupert t , strocov v n , lin h , jia s , and hasan m z 2015 _ sci . adv . _ * 1 * 1501092 xu s y , alidoust n , belopolski i , yuan z j , bian g , chang t r , zheng h , strocov v n , sanchez d s , chang g q , zhang c l , mou d x , wu y , huang l n , lee c c , huang s m , wang b k , bansil a , jeng h t , neupert t , kaminski a , lin h , jia s , and m. z. hasan 2015 _ nat . phys . _ * 11 * 748 xu n , weng h m , lv b q , matt c e , park j , bisti f , strocov v n , gawryluk d , pomjakushina e , conder k , plumb n c , radovic m , autes g , yazyev o v , fang z , dai x , qian t , mesot j , ding h , and shi m 2016 _ nat * 7 * 11006 zhang c l , xu s y , belopolski i , yuan z j , lin z q , tong b b , alidoust n , lee c c , huang s m , chang t r , jeng h t , lin h , neupane m , sanchez d s , zheng h , bian g , wang j f , zhang c , lu h z , shen s q , neupert t , hasan m z , and jia s 2016 _ nat . * 7 * 10735 kikugawa n , goswami p , kiswandhi a , choi e s , graf d , baumbach r e , brooks j s , sugii k , iida y , nishio m , uji s , terashima t , rourke p m c , hussey n e , takatsu h , yonezawa s , maeno y , and balicas l 2016 _ nat . * 7 * 10903

recently , the existence of massless chiral ( weyl ) fermions has been postulated in a class of semi - metals with a non - trivial energy dispersion . these materials are now commonly dubbed weyl semi - metals . one predicted property of weyl fermions is the chiral or adler - bell - jackiw anomaly , a chirality imbalance in the presence of parallel magnetic and electric fields . in weyl semimetals , it is expected to induce a negative longitudinal magnetoresistance , the chiral magnetic effect . here , we present experimental evidence that the observation of the chiral magnetic effect can be hindered by an effect called `` current jetting '' . this effect also leads to a strong apparent negative longitudinal magnetoresistance , but it is characterized by a highly non - uniform current distribution inside the sample . it appears in materials possessing a large field - induced anisotropy of the resistivity tensor , such as almost compensated high - mobility semimetals due to the orbital effect . in case of a non - homogeneous current injection , the potential distribution is strongly distorted in the sample . as a consequence , an experimentally measured potential difference is not proportional to the intrinsic resistance . our results on the magnetoresistance of the weyl semimetal candidate materials nbp , nbas , taas , and tap exhibit distinct signatures of an inhomogeneous current distribution , such as a field - induced `` zero resistance '' and a strong dependence of the `` measured resistance '' on the position , shape , and type of the voltage and current contacts on the sample . a misalignment between the current and the magnetic - field directions can even induce a `` negative resistance '' . finite - element simulations of the potential distribution inside the sample , using typical resistance anisotropies , are in good agreement with the experimental findings . our study demonstrates that great care must be taken before interpreting measurements of a negative longitudinal magnetoresistance as evidence for the chiral anomaly in putative weyl semimetals .

the rise of the proton structure function @xmath9 towards small bjorken @xmath3 has been discussed since the existence of qcd . in the double asymptotic limit ( large energies , i.e. small @xmath3 , and large photon virtualities @xmath2 ) the dglap evolution equations @xcite can be solved @xcite and @xmath9 is expected to rise approximately like a power of @xmath3 towards low @xmath3 . a power like behaviour is also expected in the bfkl approach @xcite . however , it soon was discussed @xcite that this rise may eventually be limited by gluon self interactions in the nucleon , or more generally due to unitarity constraints . experimentally this rise towards small @xmath3 was first observed in 1993 in the hera data @xcite . meanwhile the precision of the @xmath9 data is much improved and the rise can be studied in great detail . the low @xmath3 behaviour of @xmath9 at fixed @xmath2 is studied locally by the measurement of the derivative @xmath10 as function of @xmath3 and @xmath2 . the results are based on preliminary h1 @xmath9 data presented to this conference @xcite covering the range @xmath11 gev@xmath5 and published h1 data @xcite , @xcite which cover the range @xmath12 gev@xmath5 . the low @xmath2 @xmath9 data were obtained by shifting the @xmath13 interaction vertex by 70 cm in proton beam direction @xcite . at @xmath14 gev@xmath5 the h1 data are also shown combined with data of nmc @xcite and zeus @xcite . the derivative @xmath15 is evaluated using data points at adjacent values of @xmath3 at fixed @xmath2 taking into account error correlations and @xmath3 spacing corrections . the derivatives are compared with the next to leading order ( nlo ) qcd fit to the h1 cross section data @xcite and a `` fractal '' fit @xcite where self - similar properties of the proton structure are assumed . the @xmath3 and @xmath2 dependence of @xmath16 is shown in fig . [ fig : der1 ] . ( 200,336 ) ( -15,220 . ) ( -15,5 . ) the new shifted vertex and the published data agree well in the overlap region . the derivative @xmath17 is constant within experimental uncertainties for fixed @xmath18 in the range @xmath19 , implying that the data are consistent with the power behaviour @xmath20 . fitting this form for each @xmath2 bin to the data at @xmath19 , results in the @xmath17 and @xmath21 values presented in fig . [ fig : lamh1 ] . ( 200,315 ) ( -25,105 . ) ( -25,0 . ) the results show that the @xmath22 data at low @xmath3 for @xmath23 gev@xmath5 can be well described by the very simple parameterisation @xmath24\ ] ] with @xmath25 and @xmath26 mev and @xmath27 . at low @xmath2 the deviation of @xmath17 from the logarithmic @xmath2 dependence and the decrease of @xmath28 is more significant if the h1 data are combined with nmc @xcite and zeus @xcite data ( see fig . [ fig : lammerge ] ) . ( 200,315 ) ( -35,105 . ) ( -35,0 . ) the deviations from a simple constant respectively logarithmic behaviour occur at about such @xmath2 values below which perturbative qcd fits ( e.g. @xcite ) are not supposed to be valid . at small @xmath2 the structure function @xmath9 can be related to the total virtual photon absorption cross section by @xmath29 where the total @xmath30 energy squared is given by @xmath31 . for @xmath32 we can expect @xmath33 and @xmath34 . the latter value corresponds to the energy dependence of soft hadronic interactions @xmath35 with @xmath36 @xcite which is approximately reached at @xmath37 gev@xmath5 . no significant deviation from the power behaviour @xmath38 at fixed @xmath2 is visible at present energies and @xmath39 gev@xmath40 . more specifically : * for @xmath19 the derivative @xmath10 is independent of @xmath3 within errors . * @xmath17 is proportional to @xmath41 for @xmath42 gev@xmath5 , i.e. in the pqcd region . * here the data can be very simply parametrised by @xmath43 . * at @xmath44 gev@xmath5 deviations from the logarithmic @xmath2 dependence of @xmath17 are observed . * at low @xmath45 gev@xmath46 the energy rise is similar as in soft hadronic interactions . 99 y. l. dokshitzer , ( in russian ) , sov . jetp * 46 * ( 1977 ) 641 [ zh . * 73 * ( 1977 ) 1216 ] ; + v. n. gribov and l. n. lipatov , yad . fiz . * 15 * ( 1972 ) 1218 [ sov . j. nucl . phys . * 15 * ( 1972 ) 675 ] ; yad . fiz . * 15 * ( 1972 ) 781 [ sov . j. nucl . * 15 * ( 1972 ) 438 ] ; + g. altarelli and g. parisi , nucl . b * 126 * ( 1977 ) 298 . de rujula et al , phys . d * 10 * ( 1974 ) 1649 ; + r. d. ball and s. forte , phys . b * 335 * ( 1994 ) 77 . e. a. kuraev , l. n. lipatov and v. s. fadin , sov . jetp * 44 * ( 1976 ) 443 [ zh . * 71 * ( 1976 ) 840 ] ; sov . jetp * 45 * ( 1977 ) 199 [ zh . * 72 * ( 1977 ) 377 ] ; + i. i. balitsky and l. n. lipatov , sov . j. nucl . * 28 * ( 1978 ) 822 [ yad . fiz . * 28 * ( 1978 ) 1597 ] . l. v. gribov , e. m. levin and m. g. ryskin , nucl . b * 188 * ( 1981 ) 555 ; phys . * 100 * ( 1983 ) 1 ; + a. h. mueller and j. w. qiu , nucl . phys . b * 268 * ( 1986 ) 427 . i. abt _ et al . _ [ h1 collaboration ] , nucl . b * 407 * ( 1993 ) 515 ; + m. derrick _ et al . _ [ zeus collaboration ] , phys . b * 316 * ( 1993 ) 412 . t. lastovicka , these proceedings . m. arneodo _ et al . _ [ new muon collaboration . ] , phys . b * 364 * ( 1995 ) 107 ; nucl . b * 483 * ( 1997 ) 3 . j. breitweg _ et al . _ [ zeus collaboration ] , phys . b * 487 * ( 2000 ) 53 . t. lastovicka , arxiv : hep - ph/0203260 and these proceedings . a. donnachie and p. v. landshoff , phys . b * 296 * ( 1992 ) 227 .

results on the derivative of @xmath0 with respect to @xmath1 at fixed @xmath2 are presented . the measured derivatives are within errors independent of @xmath3 for @xmath4 gev@xmath5 and increase linearly with @xmath6 for @xmath7 and @xmath8 gev@xmath5 . the results are based on preliminary and published h1 data which at @xmath2 below 2 gev@xmath5 are combined with nmc and zeus data .

recent progress on synthetic spin - orbit coupling for neutral atoms @xcite provides new ways to study its effects in both bosonic @xcite and fermionic @xcite systems . for atomic fermi gases , many theoretical studies have shown that the spin - orbit coupling has nontrivial consequences on the bcs - bec crossover when an attractive interaction is present @xcite . even when the attraction is small , the bcs - bec crossover can be induced by increasing the spin - orbit coupling strength . due to the absence of galilean invariance in the spin - orbit coupled gases , fermionic superfluids can exhibit highly non - trivial properties @xcite . these systems can undergo quantum phase transitions to some topological superfluid phases @xcite , and it has also been shown that the finite momentum pairing state or the so - called fflo state can be energetically favored in spin - orbit coupled fermi gases @xcite . however , these more exotic phenomena require an additional , carefully engineered zeeman field . in this work , we study the response of a fermi gas with pure rashba spin - orbit coupling to a small momentum kick , in two spatial dimensions for concreteness . we find that , even in the low - momentum regime , the fermi sea is deformed in a non - trivial way , is highly degenerate , and that its energy is bound from above by that of trivial galilean - like boosts . we then consider repulsive interactions . unlike the bosonic case , where an energetic instability can be induced by a momentum kick @xcite , we find , using mean - field theory , that the ground state acquires a finite momentum for interaction strengths beyond a critical value . the fermi sea is consequently deformed , without the need to include any external zeeman fields . we also describe realistic experimental signatures of these effects in systems of ultracold atoms with the use of standard techniques . the single - particle hamiltonian of the system is given by @xcite @xmath0 where @xmath1 is the vector of spin-@xmath2 pauli matrices . the constants of motion are the helicity @xmath3 , with its corresponding eigenvectors @xmath4 forming the helicity basis , with eigenvalues @xmath5 for momentum @xmath6 , and the momentum itself . the common eigenstates of the hamiltonian , helicity and momentum have the form @xmath7,\label{singleparticlestates}\ ] ] where @xmath8 is the polar angle of @xmath9 , given by @xmath10 . the two energy branches of the system , corresponding to negative ( lower branch ) and positive ( upper branch ) helicity , respectively , are given by @xmath11 for the sake of simplicity , we will assume that the fermi gas is dilute enough that in the ground state particles only occupy negative helicity states . for this to hold , the fermi energy must be @xmath12 , with the fermi momentum @xmath13 . defining the only dimensionless parameter of the gas @xmath14 , the single - branch condition reads @xmath15 . the fermi sea in two dimensions has the form of a concentric annulus . for convenience , we define `` radii '' @xmath16 and @xmath17 for the inner and outer circumferences , respectively , which are the borders of the fermi sea . the ground state energy density @xmath18 is given by @xmath19 we study now the properties of the lowest energy fermi sea when the gas is given an infinitesimally small momentum kick . the momentum kick per particle is denoted as @xmath20 . without loss of generality , we choose the momentum kick in the positive @xmath21-direction , @xmath22 , with @xmath23 . we begin by considering the obvious galilean - like boost , consisting of assigning an extra momentum @xmath24 to each fermion in the fermi sea . we note that , formally , these are not galilean boosts since they also involve a non - trivial spin rotation , as is observed from eq . ( [ singleparticlestates ] ) . geometrically , these transformations correspond to displacing the fermi sea to the right by an amount @xmath24 . the excess energy density @xmath25 is calculated by replacing @xmath26 in the single - particle dispersion , eq . ( [ singleparticleenergy ] ) , as @xmath27 before performing the integration , we expand the modulus @xmath28 in the single - particle dispersion to second order in @xmath24 , given by @xmath29 inserting the expansion in eq . ( [ energygalileanboost1 ] ) , we obtain the excess energy for galilean - like boosts to @xmath30 , @xmath31 below , we will see that the energy of these transformations is an upper bound for the ground state energy at finite momentum , and is only attained for a saturated lower branch ( @xmath32 ) . if the momentum kick per particle @xmath20 to the gas is small , the ground state in the thermodynamic limit can be modelled by infinitesimal transformations , which describe the response of the system to small momentum kicks . these must fulfill two conditions : ( i ) the density @xmath33 is preserved and ( ii ) the momentum per particle of the resulting fermi sea equals @xmath20 . if we consider infinitesimally small momenta , we can safely rule out breaking the fermi sea into disjoint pieces . we are then left with three possibilities , namely displacements of the inner and outer circumferences and multipolar deformations of these and @xmath34 by factors @xmath35 and @xmath36 , respectively . however , these only increase the energy without changing the momentum per particle . ] we now consider displacements of the inner and outer circumferences with respect to each other , which always preserve the density of the system . the resulting momentum density is easy to calculate if we regard the empty region @xmath37 as a virtual fermi sea of holes with density @xmath38 , and the partially filled region @xmath39 as a virtual fermi sea of particles with density @xmath40 , which are defined as @xmath41 where @xmath42 . if we displace the inner and outer circumferences by momenta @xmath43 and @xmath44 , respectively , the momentum density of the system is given by @xmath45 finally , deformations of the fermi sea can be parametrized in polar coordinates by allowing the radii become angle - dependent as @xmath46 and @xmath47 , with @xmath48 and @xmath49 real periodic functions . clearly , not all deformations preserve the density . this is given by @xmath50 , which is computed from eq . [ rhoj ] with @xmath51 . this implies the following condition @xmath52 ^ 2-\left[f_i(\phi)\right]^2\right\}=0,\label{norms}\ ] ] or , equivalently @xmath53 , with @xmath54 the norm on @xmath55 . this condition is very relaxed , since it allows very different deformations of the two circumferences provided that their norms are equal . since , by convention , we have chosen to imprint a momentum to the system in the positive @xmath21-direction , we can restrict ourselves to deformations in the horizontal axis . therefore , we parametrize the functions @xmath56 ( @xmath57 ) by a multipolar expansion of the form @xmath58 where every coefficient @xmath59 is real . the condition on the norms takes the form @xmath60 . the momentum density of the deformed fermi sea is given by @xmath61 , with @xmath62 for an infinitesimally small momentum kick , we can restrict the expansion in eq . ( [ multipolar ] ) to lowest ( dipolar ) order , in which case condition ( [ norms ] ) reduces to @xmath63 . defining @xmath64 , with the upper ( lower ) sign corresponding to equal ( opposite ) dipolar coefficients , we have @xmath65 , while for @xmath66 we have @xmath67 to lowest order in @xmath24 , @xmath68 is given by @xmath69 . intuitively , dipolar coefficients with opposite sign will yield a lower energy , since a weaker deformation ( @xmath70 ) is needed in order to have the required momentum @xmath24 . the excess energy density @xmath71 for dipolar deformations is readily calculated , to leading order in @xmath72 , giving @xmath73 from the above relation , we see that for equal deformations the excess energy is that of galilean - like boosts , as in eq . ( [ energygalileanboost ] ) , @xmath74 . for opposite dipolar coefficients we obtain a lower energy given by @xmath75 ^ 2\rho,\label{energydeformations}\ ] ] which is bound from above by @xmath25 . the excess energy due to pure deformations can be lowered further by considering either displacements or both displacements and deformations . we will show next that the resulting ground - state at small non - zero momentum is highly degenerate . we now study the energy of the fermi gas when we impose displacements @xmath43 and @xmath44 of the inner and outer circumferences , together with deformations of opposite coefficients , parametrized by @xmath76 . the energy density of the system is calculated as the difference between the energy of the virtual fermi sea of particles , @xmath77 , and that of the virtual fermi sea of holes , @xmath78 , where @xmath79 with @xmath57 . to lowest order in the momenta , we can isolate the excess energy from the energy at zero momentum , which takes the form @xmath80\nonumber \\ & + 2\pi\lambda k_fq_1 ^ 2.\label{excessenergygeneral}\end{aligned}\ ] ] if the momentum per particle is set to @xmath81 , the three momenta of the system must fulfill the condition @xmath82 using the momentum constrain above to clear @xmath83 in favor of @xmath84 and @xmath85 , and inserting this in to eq . ( [ excessenergygeneral ] ) , we obtain an expression for the excess energy density , which is minimized for any @xmath84 satisfying @xmath86 with the only constrain that both @xmath84 and @xmath85 have to be small ( @xmath87 ) for our theory to be valid . we thus see that the ground - state at fixed finite momenta is infinitely degenerate . finally , the excess energy density is given by @xmath88 which is bound from above by @xmath25 for all @xmath89 and by @xmath90 for each @xmath89 . we consider now the effect of weak interactions , which we model with a zero - range potential . at mean - field level , the dirac delta needs not be regularized , and has the form @xmath91 , where @xmath92 is the interparticle distance . the dimensionless interaction parameter @xmath93 @xcite must here be compared with the only dimensionless parameter @xmath89 of the non - interacting gas . the typical interaction energy @xmath94 @xcite should be smaller than the fermi energy @xmath95 , which gives @xmath96 . the hartree shift @xmath97 is given by @xcite @xmath98 the first term is independent of the shape of the fermi sea , and the second term vanishes identically in the non - interacting ground state . however , the second term reduces the interaction energy for repulsive interactions ( @xmath99 ) if the fermi sea is modified to acquire a non - zero total momentum . without loss of generality , we again consider a momentum per particle @xmath20 in the @xmath21-direction , in which case the extra contribution to the hartree shift reads @xmath100 ^ 2.\label{extrahartree}\end{aligned}\ ] ] in the following , we minimize the total energy excess @xmath101 at a constant , infinitesimaly small momentum @xmath20 , with respect to the displacement parameter @xmath44 . we consider displacements , parametrized by @xmath83 and @xmath84 , together with deformations , parametrized by @xmath102 . the fermi sea is the set @xmath103 of momenta @xmath9 defined as @xmath104 , with @xmath105 ( @xmath106 ) . the extra hartree shift of eq . ( [ extrahartree ] ) is easily calculated and reads @xmath107 ^ 2,\label{previoushartree}\ ] ] where @xmath84 and @xmath83 are related via eq . ( [ condition ] ) , which is now valid to lowest order in @xmath24 . the energy excess @xmath108 is minimized at the point @xmath109)+q_1(2-z-2z[z+\xi])}{2(-1 + 2z^2+z\xi[1-z^2])},\label{point}\ ] ] and its value is @xmath110 where @xmath111 note that @xmath112 is independent of choice of deformation parameter @xmath85 , which shows the infinite degeneracy at small momenta holds even in the presence of interactions . as we infer from eq . ( [ finalhartree ] ) , the system s ground state , to first order , will change from the non - interacting fermi sea to a fermi sea with an infinitesimally small momentum which denotes continuity at a critical value @xmath113 where @xmath114 . this is given by @xmath115 it is important to note that , while we are considering the system s non - interacting energy to order @xmath30 , the interaction energy is not approximate but is indeed proportional to @xmath116 in the cases @xmath117 with @xmath118 and @xmath119 with @xmath120 , and therefore the above critical value is exact at the mean - field level . the critical value we have obtained is beyond the mean - field regime . this is a typical situation that also occurs , for instance , in the study of the ferromagnetic transition @xcite in repulsive fermi gases @xcite , and is resolved by going beyond first - order perturbation theory , which can result in an apparent first - order transition @xcite , or using non - perturbative methods @xcite , which predict a second - order phase transition and are in good agreement with monte - carlo simulations @xcite . our findings are the natural starting point for higher - order corrections @xcite , non - perturbative treatments , and can still be improved at the mean - field level @xcite . moreover , if , as our results suggest , the transition to finite momentum is continuous , the critical point can be obtained by calculating only the second - order response function @xmath112 but with improved treatments of the interactions . the low - momentum theory we have presented only predicts whether the system evolves towards a finite momentum ground state in a continuous manner . this is to say that our theory can describe , correctly , only derivatives in the energy at zero momenta @xmath121 the first derivative of the energy with respect to @xmath24 vanishes always at @xmath122 , while its second derivative @xmath123 for @xmath124 and @xmath125 for @xmath126 ( strictly speaking , until its pole at @xmath127 , which is far beyond the limit of validity of the mean - field theory ) . the change of sign in @xmath112 denotes the transition from a minimum to a maximum at @xmath122 , which implies that for repulsions stronger than the critical value @xmath128 the system s ground state has non - zero momentum . to calculate the actual momentum of the ground state for @xmath129 it is necessary to go beyond leading order in @xmath24 . so far , we have considered the momentum kick of the gas to be given in a particular direction . in an experiment , however , no direction is in principle preferred . obviously , after the critical point is reached , the ground - state is highly degenerate , since the same momentum in any direction yields the same energy . the many - body wave function is therefore an arbitrary superposition of states with momenta pointing at different directions . for instance , we may expect to observe an equal superposition of these states , up to arbitrary phases , which yields a circularly symmetric momentum distribution . momentum distributions are especially relevant to cold atom experiments where these can be obtained via time - of - flight measurements @xcite , in combination with spin - injection spectroscopy @xcite or momentum resolved rf spectroscopy @xcite . ( color online ) integrated momentum distribution ( black solid line ) at critical interaction strength @xmath130 , with @xmath131 and @xmath132 , compared to the non - interacting momentum distribution ( red dashed line ) . ] we can easily map out , starting from a fermi sea at finite momentum in a particular direction , the aforementioned circularly symmetric momentum distribution . the momentum distribution is denoted by @xmath133 , and is such that @xmath134 . we define the integrated , angle - independent momentum distribution @xmath135 as @xmath136 geometrically , @xmath135 is the arc length of occupied states on a circumference of radius @xmath137 in momentum space . closed circumferences map into unit length fully occupied states while open arcs give shorter lengths smaller average occupations . the circularly symmetric momentum distribution arising from the superposition of the different states is obtained as a surface of revolution by rotating the fermi sea at finite momentum in a particular direction with respect to the origin in momentum space . clearly , the resulting momentum distribution coincides with @xmath135 . in fig . [ ref ] , we show the integrated momentum distribution for a particular case of an interacting fermi sea at non - zero total momentum compared to the non - interacting momentum distribution . there , we observe a shortening of the unit occupation plateau , together with a smoothening of the momentum distribution at the edges of the fermi sea , with an obvious change in concavity , which can be a relevant experimental signature for finite momentum states . in cold atom experiments , where an external trap is always present , the fermi sea and the homogeneous momentum distribution can be observed by selectively probing fermions around the centre of the trap @xcite . an alternative way to observe these effects consists of adding to the system a small symmetry - breaking term , i.e. a small momentum kick in a chosen direction , in order to observe the deformations _ per se_. this can routinely be done nowadays with the use of standing - wave light - pulse sequences @xcite . this technique has been successfully applied to ultracold atom systems subject to artificial spin - orbit coupling @xcite . in this paper , we have studied the response of a dilute two - dimensional fermi gas with rashba spin - orbit coupling to a small overall constant velocity kick . we have found that the moving fermi sea deforms in a non - trivial manner due to the non - galilean nature of the system , and is highly degenerate . we have then considered repulsive interactions at the hartree - fock level , and found that the ground - state of the system acquires a finite - momentum . the fermi sea becomes deformed beyond a critical interaction strength in a continuous fashion , which we identified as a possible experimental signature . our results open the path towards the observation of finite momentum ground states , constitute the starting point for more elaborate treatments of interactions , and can be generalized to higher dimensions and more general types of spin - orbit coupling . 99 j. dalibard , f. gerbier , g. juzelinas and p. hberg , rev . phys . * 83 * , 1523 ( 2011 ) j. d. sau , r. sensanrma , s. powell , i. b. spielman and s. das sarma , phys . rev . b * 83 * , 140510(r ) ( 2011 ) d. l. campbell , g. juzelinas and i. b. spielman , phys . rev . a * 84 * , 025602 ( 2011 ) g. juzelinas , j. ruseckas and j. dalibard , phys . a * 81 * , 053403 ( 2010 ) . x. j. liu , m. f. borunda , x. liu and j. sinova , phys . . lett . * 102 * , 046402 ( 2009 ) y. j. lin , r. l. compton , k. jimnez - garca , j. v. porto and i. b. spielman , nature ( london ) * 462 * , 628 ( 2009 ) y. j. lin , k. jimnez - garca and i. b. spielman , nature ( london ) * 471 * , 83 ( 2011 ) . p. wang , z .- q . yu , z. fu , j. miao , l. huang , s. chai , h. zhai and j. zhang , phys . lett . * 109 * , 095301 ( 2012 ) l. w. cheuk , a. t. sommer , z. hadzibabic , t. yefsah , w. s. bakr and m. w. zwierlein , phys . lett . * 109 * , 095302 ( 2012 ) . j. p. vyasanakere , s. zhang and v. b. shenoy , phys . rev.b * 84 * , 014512 ( 2011 ) h. hu , l. liang , x .- j . liu and h. pu , phys . lett . * 107 * , 195304 ( 2011 ) z .- q . yu and h. zhai , phys . lett . * 107 * , 195305 ( 2011 ) l. he and x .- huang , phys . . lett . * 108 * , 145302 ( 2012 ) l. han and c. a. r. s de melo , phys . a * 85 * , 011606(r)(2012 ) . m. gong , s. tewari and c. zhang , phys . 107 * , 195303 ( 2011 ) m. iskin and a. l. subasi , phys . lett . * 107 * , 050402 ( 2011 ) ; e - print arxiv : 1211.4020 k. seo , l. han and c. a. r. s de melo , phys . rev . a * 85 * , 033601 ( 2012 ) f. wu , g .- c . guo , w. zhang and w. wi , e - print arxiv:1211.5780 . z. zheng , m. gong , x. zou , c. zhang and g. guo , e - print arxiv : 1208.2029 z. zheng , m. gong , y. zhang , x. zou , c. zhang and g. cuo , e - print arxiv : 1212.6826 l. dong , l. jiang , h. hu and h. pu , e - print arxiv:1211.1700 v. b. shenoy , e - print arxiv:1211.1831 x .- j . liu and h. hu , e - print arxiv:1302.0553 . r. a. williams , l. j. leblanc , k. jimnez - garca , m. c. beeler , a. r. perry , w. d. phillips and i. b. spielman , science * 335 * , 314 ( 2012 ) ; see supporting online material at http://www.sciencemag.org/content/suppl/2011/12/08/science.1212652.dc1.html .

we consider the effect of a momentum kick on the ground state of a non - interacting two - dimensional fermi gas subject to rashba spin - orbit coupling . although the total momentum is a constant of motion , the gas does not obey the rules of galilean relativity . upon imprinting a small overall velocity to the non - interacting gas , we find that the fermi sea is deformed in a non - trivial way . we also consider a weakly repulsive fermi gas , and find , from its hartree shift , that the total ground state of the system may change into a deformed , finite momentum ground state as the repulsion is increased beyond a critical value , without the need of any external zeeman fields . we also discuss possible experimental signatures of these effects .

although the top quark was discovered almost ten years ago @xcite , many of its properties are still only poorly known @xcite . in particular , the couplings of the top quark to the electroweak ( ew ) gauge bosons have not yet been directly measured . current data provide only weak constraints on the couplings of the top quark with the ew gauge bosons , except for the @xmath3 vector and axial vector couplings which are rather tightly but indirectly constrained by lep data ; and the right - handed @xmath4 coupling , which is severely bound by the observed @xmath5 rate @xcite . at an @xmath6 linear collider with @xmath7 gev and an integrated luminosity of @xmath8 fb@xmath9 one can hope to measure the @xmath10 ( @xmath11 ) couplings in top pair production with a few - percent precision @xcite . however , the process @xmath12 is sensitive to both @xmath13 and @xmath3 couplings and significant cancellations between the various couplings can occur . at hadron colliders , @xmath14 production is so dominated by the qcd processes @xmath15 and @xmath16 that a measurement of the @xmath13 and @xmath3 couplings via @xmath17 is hopeless . instead , the @xmath10 couplings can be measured in qcd @xmath18 production , radiative top quark decays in @xmath14 events ( @xmath19 ) , and qcd @xmath20 production @xcite . @xmath18 production and radiative top quark decays are sensitive only to the @xmath13 couplings , whereas @xmath20 production gives information only on the structure of the @xmath3 vertex . this obviates having to disentangle potential cancellations between the different couplings . in this contribution we briefly review the measurement of the @xmath10 couplings at the lhc and compare the expected sensitivities with the bounds which one hopes to achieve at an @xmath6 linear collider . the most general lorentz - invariant vertex function describing the interaction of a neutral vector boson @xmath21 with two top quarks can be written in terms of ten form factors @xcite , which are functions of the kinematic invariants . in the low energy limit , these correspond to couplings which multiply dimension - four or -five operators in an effective lagrangian , and may be complex . if @xmath21 is on - shell , or if @xmath21 couples to effectively massless fermions , the number of independent form factors is reduced to eight . if , in addition , both top quarks are on - shell , the number is further reduced to four . in this case , the @xmath10 vertex can be written in the form @xmath22 where @xmath23 is the proton charge , @xmath24 is the top quark mass , @xmath25 is the outgoing top ( anti - top ) quark four - momentum , and @xmath26 . the terms @xmath27 and @xmath28 in the low energy limit are the @xmath10 vector and axial vector form factors . the coefficients @xmath29 and @xmath30 are related to the magnetic and ( @xmath31-violating ) electric dipole form factors . in @xmath32 production , one of the top quarks coupling to @xmath21 is off - shell . the most general vertex function relevant for @xmath32 production thus contains additional couplings , not included in eq . ( [ eq : anomvertex ] ) . these additional couplings are irrelevant in @xmath33 , where both top quarks are on - shell . in @xmath33 one often uses the following parameterization for the @xmath10 vertex : @xmath34 using the gordon decomposition , it is easy to show that eqs . ( [ eq : anomvertex ] ) and ( [ eq : gordon ] ) are equivalent for onshell top quarks and that the form factors @xmath35 and @xmath36 ( @xmath37 ) are related by @xmath38 the process @xmath39 followed by @xmath40 leads either to a @xmath41 final state if both @xmath42 bosons decay leptonically , to a @xmath43 final state if one @xmath42 decays leptonically and the other decays hadronically , or to a @xmath44 final state if both @xmath42 bosons decay hadronically . the @xmath44 final state has the largest br . however , it is plagued by a large qcd background , so we ignore it . the dilepton final state , although less contaminated by background , has a br about a factor 6 smaller than that of the so - called lepton+jets mode . in the following , we therefore concentrate on @xmath45 with @xmath46 . we assume that both @xmath47 quarks are tagged with a combined efficiency of @xmath48 . we perform our calculation for general @xmath13 couplings of the form of eq . ( [ eq : anomvertex ] ) . we otherwise assume the sm to be valid . our calculation includes top quark and @xmath42 decays with full spin correlations and finite width effects . all resonant feynman diagrams contributing to the lepton+jets final state are included , i.e. besides @xmath18 production , we automatically take into account top quark pair production where one of the top quarks decays radiatively , @xmath49 . we impose standard acceptance cuts for leptons , jets , and the missing transverse momentum . a detailed list can be found in ref . we also include minimal detector effects via gaussian smearing of parton momenta according to cms @xcite expectations , and take into account the @xmath47 jet energy loss via a parameterized function . since we are interested in photon emission from top quarks , we would like to suppress radiation from @xmath42 decay products , as well as emission from @xmath47 quarks . imposing a large separation cut of @xmath50 reduces photon radiation from the @xmath47 quarks . photon emission from @xmath42 decay products can essentially be eliminated by requiring that @xmath51 and @xmath52 where @xmath53 is the invariant mass of the @xmath54 system , and @xmath55 is the @xmath56 cluster transverse mass , which peaks sharply at @xmath57 . in addition we require that the event is consistent either with @xmath18 production , or with @xmath14 production with radiative top decay . this will reduce the singly - resonant and non - resonant backgrounds . the invariant mass and cluster transverse mass cuts which are imposed to accomplish this can be found in ref . @xcite . the most important irreducible background processes that remain after imposing the cuts described above , are single - top processes ( @xmath58 ) , and the non - resonant process @xmath59 . we calculate the irreducible background processes at leading order in qcd including the full set of contributing feynman diagrams using madevent @xcite . the potentially most dangerous reducible background is @xmath60 production where one of the jets in the final state fakes a photon . [ cols="^,^ " , ] [ tab : tab1 ] it is instructive to compare the bounds for anomalous @xmath10 couplings achievable at the lhc with those projected for the ilc . the most complete study of @xmath14 production at the ilc for general @xmath10 ( @xmath11 ) couplings so far is that of ref . it uses the parameterization of eq . ( [ eq : gordon ] ) for the @xmath10 vertex function . in order to compare the bounds of ref . @xcite with those anticipated at the lhc , the limits on @xmath61 and @xmath62 have to be converted into bounds on @xmath63 and @xmath64 . table [ tab : tab1 ] compares the bounds we obtain for @xmath65 and @xmath66 with those reported for the ilc in ref . we show lhc limits only for an integrated luminosity of 300 fb@xmath9 . the results of table [ tab : tab1 ] demonstrate that the ilc , with the exception of @xmath67 and @xmath68 , will be able to considerably improve the sensitivity limits which can be achieved at the lhc , in particular for the @xmath3 couplings . the lhc will be able to perform first tests of the @xmath10 couplings . already with an integrated luminosity of 30 fb@xmath9 , one can probe the @xmath13 couplings with a precision of about @xmath69 per experiment . with higher integrated luminosities one will be able to reach the few percent region . the @xmath20 cross section with leptonic @xmath70 decays is roughly a factor 20 smaller than the @xmath18 rate . it is therefore not surprising that the sensitivity limits on the @xmath3 couplings are significantly weaker than those which one expects for the @xmath13 couplings . the ilc , with the exception of @xmath67 and @xmath68 , will be able to further improve our knowledge of the @xmath10 couplings , in particular in the @xmath3 case . 99 f. abe _ et al . _ ( cdf collaboration ) , phys . lett . * 74 * , 2626 ( 1995 ) . s. abachi _ et al . _ ( d collaboration ) , phys . lett . * 74 * , 2632 ( 1995 ) . d. chakraborty , j. konigsberg and d. l. rainwater , ann . rev . nucl . sci . * 53 * , 301 ( 2003 ) . f. larios , m. a. perez and c. p. yuan , phys . b457 * , 334 ( 1999 ) ; m. frigeni and r. rattazzi , phys . lett . * b269 * , 412 ( 1991 ) . t. abe _ et al . _ ( american linear collider working group collaboration ) , arxiv : hep - ex/0106057 . u. baur , a. juste , l. h. orr and d. rainwater , phys . rev . * d71 * , 054013 ( 2005 ) . w. hollik _ et al . _ , nucl . phys . * b551 * , 3 ( 1999 ) [ erratum - ibid . * b557 * , 407 ( 1999 ) ] .

the international linear collider ( ilc ) will be able to precisely measure the electroweak couplings of the top in @xmath0 . we compare the limits which can be achieved at the ilc with those which can be obtained in @xmath1 and @xmath2 production at the large hadron collider ( lhc ) .

along with turbo codes , low - density parity - check ( ldpc ) block codes form a class of codes which approach the ( theoretical ) shannon limit . ldpc codes were first introduced in the @xmath0s by gallager @xmath1 . however , they were considered impractical at that time and very little related work was done until tanner provided a graphical interpretation of the parity - check matrix in 1981 @xmath2 . more recently , in his ph.d . thesis , wiberg revived interest in ldpc codes and further developed the relation between tanner graphs and iterative decoding @xcite . the convolutional counterpart of ldpc block codes was introduced in @xcite , and ldpc convolutional codes have been shown to have certain advantages compared to ldpc block codes of the same complexity @xcite . in this paper , we use ensembles of tail - biting ldpc convolutional codes derived from a protograph - based ensemble of ldpc block codes to obtain a lower bound on the free distance of unterminated , asymptotically good , periodically time - varying ldpc convolutional code ensembles , i.e. , ensembles that have the property of free distance growing linearly with constraint length . in the process , we show that the minimum distances of ensembles of tail - biting ldpc convolutional codes ( introduced in @xcite ) approach the free distance of an associated unterminated , periodically time - varying ldpc convolutional code ensemble as the block length of the tail - biting ensemble increases . we also show that , for protographs with regular degree distributions , the free distance bounds are consistent with those recently derived for regular ldpc convolutional code ensembles in @xcite and @xcite . further , for protographs with irregular degree distributions , we obtain new free distance bounds that grow linearly with constraint length and whose free distance to constraint length ratio exceeds the minimum distance to block length ratio of the corresponding block codes . the paper is structured as follows . in section [ sec : ldpccc ] , we briefly introduce ldpc convolutional codes . section [ sec : proto ] summarizes the technique proposed by divsalar to analyze the asymptotic distance growth behavior of protograph - based ldpc block codes @xcite . in section [ sec : distbnd ] , we describe the construction of tail - biting ldpc convolutional codes as well as the corresponding unterminated , periodically time - varying ldpc convolutional codes . we then show that the free distance of a periodically time - varying ldpc convolutional code is lower bounded by the minimum distance of the block code formed by terminating it as a tail - biting ldpc convolutional code . finally , in section [ sec : results ] we present new results on the free distance of ensembles of ldpc convolutional codes based on protographs . we start with a brief definition of a rate @xmath3 binary ldpc convolutional code @xmath4 . ( a more detailed description can be found in @xcite . ) a code sequence @xmath5}$ ] satisfies the equation @xmath6}\mathbf{h}_{[0,\infty ] } ^{\mathtt{t}}=\mathbf{0},\ ] ] where @xmath7}^{\mathtt{t}}$ ] is the syndrome former matrix and @xmath7}=$ ] + is the parity - check matrix of the convolutional code @xmath4 . the submatrices @xmath8 , @xmath9 , @xmath10 , are binary @xmath11 submatrices , given by @xmath12,\ ] ] that satisfy the following properties : 1 . @xmath13 and @xmath14 2 . there is a @xmath15 such that @xmath16 3 . @xmath17 and has full rank @xmath18 . we call @xmath19 the syndrome former memory and @xmath20 the decoding constraint length . these parameters determine the width of the nonzero diagonal region of @xmath7}.$ ] the sparsity of the parity - check matrix is ensured by demanding that its rows have very low hamming weight , i.e. , @xmath21 , where @xmath22 denotes the @xmath23-th row of @xmath24}$ ] . the code is said to be regular if its parity - check matrix @xmath7}$ ] has exactly @xmath25 ones in every column and , starting from row @xmath26 , @xmath27 ones in every row . the other entries are zeros . we refer to a code with these properties as an @xmath28-regular ldpc convolutional code , and we note that , in general , the code is time - varying and has rate @xmath29 . an @xmath28-regular time - varying ldpc convolutional code is periodic with period @xmath30 if @xmath8 is periodic , i.e. , @xmath31 , and if @xmath32 , the code is time - invariant . an ldpc convolutional code is called irregular if its row and column weights are not constant . the notion of degree distribution is used to characterize the variations of check and variable node degrees in the tanner graph corresponding to an ldpc convolutional code . optimized degree distributions have been used to design ldpc convolutional codes with good iterative decoding performance in the literature ( see , e.g. , @xcite ) , but no distance bounds for irregular ldpc convolutional code ensembles have been previously published . suppose a given protograph has @xmath33 variable nodes and @xmath34 check nodes . an ensemble of protograph - based ldpc block codes can be created by the copy - and - permute operation @xcite . the tanner graph obtained for one member of an ensemble created using this method is illustrated in fig . [ fig : proto ] . the parity - check matrix @xmath35 corresponding to the ensemble of protograph - based ldpc block codes can be obtained by replacing ones with @xmath36 permutation matrices and zeros with @xmath36 all zero matrices in the underlying protograph parity - check matrix @xmath37 , where the permutation matrices are chosen randomly and independently . the protograph parity - check matrix @xmath37 corresponding to the protograph given in figure [ fig : proto ] can be written as where we note that , since the row and column weights of @xmath37 are not constant , @xmath37 represents the parity - check matrix of an irregular ldpc code . if a variable node and a check node in the protograph are connected by @xmath38 parallel edges , then the associated entry in @xmath37 equals @xmath38 and the corresponding block of @xmath35 consists of a summation of @xmath38 @xmath36 permutation matrices . the sparsity condition of an ldpc parity - check matrix is thus satisfied for large @xmath39 . the code created by applying the copy - and - permute operation to an @xmath40 protograph parity - check matrix @xmath37 has block length @xmath41 . in addition , the code has the same rate and degree distribution for each of its variable and check nodes as the underlying protograph code . combinatorial methods of calculating ensemble average weight enumerators have been presented in @xcite and @xcite . the remainder of this section summarizes the methods presented in @xcite . suppose a protograph contains @xmath42 variable nodes to be transmitted over the channel and @xmath43 punctured variable nodes . also , suppose that each of the @xmath42 transmitted variable nodes has an associated weight @xmath44 , where @xmath45 for all @xmath23 . copies of the protograph , the weight associated with a particular variable node in the protograph can be as large as @xmath39 . ] let @xmath46 be the set of all possible weight distributions such that @xmath47 , and let @xmath48 be the set of all possible weight distributions for the remaining punctured nodes . the ensemble weight enumerator for the protograph is then given by @xmath49 where @xmath50 is the average number of codewords in the ensemble with a particular weight distribution @xmath51 . the normalized logarithmic asymptotic weight distribution of a code ensemble can be written as @xmath52 where @xmath53 , @xmath54 , @xmath55 is the hamming distance , @xmath56 is the block length , and @xmath57 is the ensemble average weight distribution . suppose the first zero crossing of @xmath58 occurs at @xmath59 . if @xmath58 is negative in the range @xmath60 , then @xmath61 is called the _ minimum distance growth rate _ of the code ensemble . by considering the probability @xmath62 it is clear that , as the block length @xmath56 grows , if @xmath63 , then we can say with high probability that the majority of codes in the ensemble have a minimum distance that grows linearly with @xmath56 and that the distance growth rate is @xmath61 . in this section we present a method for obtaining a lower bound on the free distance of an ensemble of unterminated , asymptotically good , periodically time - varying ldpc convolutional codes derived from protograph - based ldpc block codes . to proceed , we will make use of a family of tail - biting ldpc convolutional codes with incremental increases in block length . the tail - biting codes will be used as a tool to obtain the desired bound on the free distance of the unterminated codes . suppose that we have an @xmath40 protograph parity - check matrix @xmath37 , where gcd@xmath64 . we then partition @xmath37 as a @xmath65 block matrix as follows : @xmath66,\ ] ] where each block @xmath67 is of size @xmath68 . @xmath37 can thus be separated into a lower triangular part , @xmath69 , and an upper triangular part minus the leading diagonal , @xmath70 . explicitly , where blank spaces correspond to zeros . this operation is called ` cutting ' a protograph parity - check matrix . rearranging the positions of these two triangular matrices and repeating them indefinitely results in a parity - check matrix @xmath71 of an unterminated , periodically time - varying convolutional code with constraint length @xmath72 and period @xmath73 given by of @xmath74 , where @xmath75 divides @xmath76 without remainder . if @xmath77 then the resulting convolutional code is time - invariant . ] @xmath78.\ ] ] note that if gcd@xmath79 , we can not form a square block matrix larger than @xmath80 with equal size blocks . in this case , @xmath81 and @xmath70 is the all zero matrix of size @xmath82 . this trivial cut results in a convolutional code with syndrome former memory zero , with repeating blocks of the original protograph on the leading diagonal . it is necessary in this case to create a larger protograph parity - check matrix by using the copy and permute operation on @xmath37 . this results in an @xmath83 parity - check matrix for some small integer @xmath84 . the @xmath85 protograph parity - check matrix can then be cut following the procedure outlined above . in effect , the choice of @xmath86 permutation matrix creates a mini ensemble of block codes suitable to be unwrapped to an ensemble of convolutional codes . we now introduce the notion of tail - biting convolutional codes by defining an ` unwrapping factor ' @xmath87 as the number of times the sliding convolutional structure is repeated . for @xmath88 , the parity - check matrix @xmath89 of the desired tail - biting convolutional code can be written as @xmath90_{\lambda n_c \times \lambda n_v}.\ ] ] note that the tail - biting convolutional code for @xmath91 is simply the original block code . given a protograph parity - check matrix @xmath37 , we generate a family of tail - biting convolutional codes with increasing block lengths @xmath92 , @xmath93 , using the process described above . since tail - biting convolutional codes are themselves block codes , we can treat the tanner graph of @xmath89 as a protograph for each value of @xmath87 . replacing the entries of this matrix with either @xmath94 permutation matrices or @xmath36 all zero matrices , as discussed in section [ sec : proto ] , creates an ensemble of ldpc codes that can be analyzed asymptotically as @xmath39 goes to infinity , where the sparsity condition of an ldpc code is satisfied for large @xmath39 . each tail - biting ldpc code ensemble , in turn , can be unwrapped and repeated indefinitely to form an ensemble of unterminated , periodically time - varying ldpc convolutional codes with constraint length @xmath95 and , in general , period @xmath96 . intuitively , as @xmath87 increases , the tail - biting code becomes a better representation of the associated unterminated convolutional code , with @xmath97 corresponding to the unterminated convolutional code itself . this is reflected in the weight enumerators , and it is shown in section [ sec : results ] that increasing @xmath87 provides us with distance growth rates that converge to a lower bound on the free distance growth rate of the unterminated convolutional code . tail - biting convolutional codes can be used to establish a lower bound on the free distance of an associated unterminated , periodically time - varying convolutional code by showing that the free distance of the unterminated code is lower bounded by the minimum distance of any of its tail - biting versions . a proof can be found in @xcite . consider a rate @xmath98 unterminated , periodically time - varying convolutional code with decoding constraint length @xmath99 and period @xmath96 . let @xmath100 be the minimum distance of the associated tail - biting convolutional code with length @xmath101 and unwrapping factor @xmath102 . then the free distance @xmath103 of the unterminated convolutional code is lower bounded by @xmath100 for any unwrapping factor @xmath87 , i.e. , @xmath104 a trivial corollary of the above theorem is that the minimum distance of a protograph - based ldpc block code is a lower bound on the free distance of the associated unterminated , periodically time - varying ldpc convolutional code . this can be observed by setting @xmath105 . one must be careful in comparing the distance growth rates of codes with different underlying structures . a fair basis for comparison generally requires equating the complexity of encoding and/or decoding of the two codes . traditionally , the minimum distance growth rate of block codes is measured relative to block length , whereas constraint length is used to measure the free distance growth rate of convolutional codes . these measures are based on the complexity of decoding both types of codes on a trellis . indeed , the typical number of states required to decode a block code on a trellis is exponential in the block length , and similarly the number of states required to decode a convolutional code is exponential in the constraint length . this has been an accepted basis of comparing block and convolutional codes for decades , since maximum - likelihood decoding can be implemented on a trellis for both types of codes . the definition of decoding complexity is different , however , for ldpc codes . the sparsity of their parity - check matrices , along with the iterative message - passing decoding algorithm typically employed , implies that the decoding complexity per symbol depends on the degree distribution of the variable and check nodes and is independent of both the block length and the constraint length . the cutting technique we described in section [ sec : tb ] preserves the degree distribution of the underlying ldpc block code , and thus the decoding complexity per symbol is the same for the block and convolutional codes considered in this paper . also , for randomly constructed ldpc block codes , state - of - the - art encoding algorithms require only @xmath106 operations per symbol , where @xmath107 @xcite , whereas for ldpc convolutional codes , if the parity - check matrix satisfies the conditions listed in section ii , the number of encoding operations per symbol is only @xmath108 @xcite . here again , the encoding complexity per symbol is essentially independent of both the block length and the constraint length . hence , to compare the distance growth rates of ldpc block and convolutional codes , we consider the hardware complexity of implementing the encoding and decoding operations in hardware . typical hardware storage requirements for both ldpc block encoders and decoders are proportional to the block length @xmath56 . the corresponding hardware storage requirements for ldpc convolutional encoders and decoders are proportional to the decoding constraint length @xcite . , encoding constraint lengths may be preferred to decoding constriant lengths . for further details , see @xcite . ] we now present distance growth rate results for several ensembles of rate @xmath109 asymptotically good ldpc convolutional codes based on protographs . _ example @xmath110 _ consider a @xmath111 regular ldpc code with the folowing protograph : for this example , the minimum distance growth rate is @xmath112 , as originally calculated by gallager @xcite . a family of tail - biting ldpc convolutional code ensembles can be generated according to the following cut : for each @xmath87 , the minimum distance growth rate @xmath61 was calculated for the tail - biting ldpc convolutional codes using the approach outlined in section [ sec : tailll ] . the distance growth rates for each @xmath87 are given as @xmath113 the free distance growth rate of the associated rate @xmath109 ensemble of unterminated , periodically time - varying ldpc convolutional codes is @xmath114 , as discussed above . then ( [ emre ] ) gives us the lower bound @xmath115 for @xmath116 . these growth rates are plotted in fig . [ fig : ex1 ] . .,width=307 ] we observe that , once the unwrapping factor @xmath87 of the tail - biting convolutional codes exceeds @xmath117 , the lower bound on @xmath118 levels off at @xmath119 , which agrees with the results presented in @xcite and @xcite and represents a significant increase over the value of @xmath61 . in this case , the minimum weight codeword in the unterminated convolutional code also appears as a codeword in the tail - biting code . example @xmath120 _ the following irregular protograph is from the repeat jagged accumulate @xcite ( rja ) family . it was shown to have a good iterative decoding threshold ( @xmath121 db ) while maintaining linear minimum distance growth ( @xmath122 ) . we display below the associated @xmath37 matrix and cut used to generate the family of tail - biting ldpc convolutional code ensembles . we observe that , as in example @xmath110 , the minimum distance growth rates calculated for increasing @xmath87 provide us with a lower bound on the free distance growth rate of the convolutional code ensemble using ( [ lb ] ) . the lower bound was calculated as @xmath123 ( for @xmath124 ) , significantly larger than the minimum distance growth rate @xmath61 of the underlying block code ensemble . _ example @xmath117 _ the following irregular protograph is from the accumulate repeat jagged accumulate family ( arja ) @xcite : where the undarkened circle represents a punctured variable node . this protograph is of significant practical interest , since it was shown to have @xmath125 and iterative decoding threshold @xmath126 , i.e. , pre - coding the protograph of example @xmath120 provides an improvement in both values . in this arja example , the protograph matrix @xmath37 is of size @xmath127 . we observe that gcd@xmath79 , and thus we have the trivial cut mentioned in section [ sec : tb ] . we must then copy and permute @xmath37 to generate a mini ensemble of block codes . results are shown for one particular member of the mini ensemble with @xmath128 , but a change in performance can be obtained by varying the particular permutation chosen . increasing @xmath87 for the chosen permutation results in a lower bound , found using ( [ lb ] ) , of @xmath129 for @xmath130 . again , we observe a significant increase in @xmath118 compared to @xmath61 . simulation results for ldpc block and convolutional codes based on the protograph of example @xmath117 were obtained assuming bpsk modulation and an additive white gaussian noise channel ( awgnc ) . all decoders were allowed a maximum of @xmath131 iterations , and the block code decoders employed a syndrome - check based stopping rule . as a result of their block structure , tail - biting ldpc convolutional codes were decoded using standard ldpc block decoders employing a belief - propagation decoding algorithm . the ldpc convolutional code , on the other hand , was decoded by a sliding - window based belief - propagation decoder @xcite . the resulting bit error rate ( ber ) performance is shown in fig.[fig : sim ] . .,width=336 ] we note that the protograph - based tail - biting ldpc convolutional codes outperform the underlying protograph - based ldpc block code ( which can also be seen as a tail - biting code with unwrapping factor @xmath132 ) . larger unwrapping factors yield improved error performance , eventually approaching the performance of the unterminated convolutional code , which can be seen as a tail - biting code with an infinitely large unwrapping factor . we also note that no error floor is observed for the convolutional code , which is expected , since the code ensemble is asymptotically good and has a relatively large ( @xmath129 ) distance growth rate . we also note that the performance of the unterminated ldpc convolutional code is consistent with the iterative decoding threshold computed for the underlying protograph . at a moderate constraint length of @xmath133 , the unterminated code achieves @xmath134 ber at roughly @xmath135 db away from the threshold , and with larger block ( constraint ) lengths , the performance will improve even further . this is expected , since both the unterminated and the tail - biting convolutional codes preserve the same degree distribution as the underlying protograph . in this paper , asymptotic methods were used to calculate a lower bound on the free distance that grows linearly with constraint length for several ensembles of unterminated , protograph - based periodically time varying ldpc convolutional codes . it was shown that the free distance growth rates of the ldpc convolutional code ensembles exceed the minimum distance growth rates of the corresponding ldpc block code ensembles . further , we observed that the performance of the ldpc convolutional codes is consistent with the iterative decoding thresholds of the underlying protographs . this work was partially supported by nsf grants ccr02 - 05310 and ccf05 - 15012 and nasa grants nng05gh736 and nnx07ak536 . in addition , the authors acknowledge the support of the scottish funding council for the joint research institute with the heriot - watt university , which is a part of the edinburgh research partnership . mitchell acknowledges the royal society of edinburgh for the award of the john moyes lessells travel scholarship . 1 r. g. gallager , `` low - density parity - check codes '' , _ ire trans . inform . theory _ , it-8 : 21 - 28 , jan . 1962 . r. michael tanner , `` a recursive approach to low complexity codes '' , _ ieee trans . inform . theory _ , it-27 , pp.533 - 547 , sept . n. wiberg , codes and decoding on general graphs " , ph.d . thesis , dept . of e. e. , linkping university , linkping , sweden . a. jimnez - felstrm and k. sh . zigangirov , `` time - varying periodic convolutional codes with low - density parity - check matrices '' , _ ieee trans . inform . theory _ , it-45 , pp.2181 - 2191 , sept d. j. costello , jr . , a. pusane , s. bates , and k. sh . zigangirov , `` a comparison between ldpc block and convolutional codes '' , _ proc . inform . theory and app . workshop _ , san diego , ca , feb . 2006 . d. j. costello , jr . , a. e. pusane , c. r. jones , and d. divsalar , `` a comparison of ara- and protograph - based ldpc block and convolutional codes '' , _ proc . inform . theory and app . workshop _ , san diego , ca , feb . m. tavares , k. sh . zigangirov , and g. fettweis , `` tail - biting ldpc convolutional codes '' , _ proc . symp . on inform . theory _ , nice , france , june 2007 . a. sridharan , d. truhachev , m. lentmaier , d. j. costello , jr . , and k. sh . zigangirov , `` distance bounds for an ensemble of ldpc convolutional codes '' , _ ieee trans . inform . theory _ , it-53 , pp . 4537 - 4555 , dec . 2007 . d. truhachev , k. sh . zigangirov , and d. j. costello , jr . , `` distance bounds for periodic ldpc convolutional codes and tail - biting convolutional codes '' , _ ieee trans . inform . theory _ ( submitted ) , jan . available online at http://luur.lub.lu.se/luur?func=downloadfile + & fileoid=1054996 . d. divsalar , `` ensemble weight enumerators for protograph ldpc codes '' , _ proc . symp . on inform . seattle , wa , jul . 2006 . g. richter , m. kaupper , and k. sh . zigangirov `` irregular low - density parity - check convolutional codes based on protographs '' , _ proc . symp . on inform . seattle , wa , jul . a. sridharan , d. sridhara , d. j. costello , jr . , and t. e. fuja `` a construction for irregular low density parity check codes '' , _ proc . symp . on inform . theory _ , yokohama , japan , june 2003 . a. e. pusane , k. sh . zigangirov , and d. j. costello , jr . , `` construction of irregular ldpc convolutional codes with fast decoding '' , _ proc . conf . on communications _ , istanbul , turkey , june 2006 . j. thorpe , `` low - density parity - check ( ldpc ) codes constructed from protographs '' , _ jpl inp progress report _ , vol . 42 - 154 , aug . s. l. fogal , r. mceliece , and j. thorpe , `` enumerators for protograph ensembles of ldpc codes '' , _ proc . symp . on inform . adelaide , australia , sept . 2005 . t. j. richardson and r. l. urbanke , `` efficient encoding of low - density parity - check codes , '' _ ieee trans . inform . theory _ , it-47 , pp.638 - 656 , feb a. e. pusane , a. jimnez - feltstrm , a. sridharan , m. lentmaier , k. sh . zigangirov , and d. j. costello , jr . , `` implementation aspects of ldpc convolutional codes , '' _ ieee trans . commun . _ , to appear . d. g. m. mitchell , a. e. pusane , n. goertz , and d. j. costello , jr . , `` free distance bounds for protograph - based regular ldpc convolutional codes '' , _ 2008 int . symp . on turbo codes and related topics _ ( submitted ) . available online at http://arxiv.org/abs/0804.4466 . d. divsalar , s. dolinar , and c. jones , `` construction of protograph ldpc codes with linear minimum distance '' , _ proc . symp . on inform . theory _ , seattle , wa , jul .

ldpc convolutional codes have been shown to be capable of achieving the same capacity - approaching performance as ldpc block codes with iterative message - passing decoding . in this paper , asymptotic methods are used to calculate a lower bound on the free distance for several ensembles of asymptotically good protograph - based ldpc convolutional codes . further , we show that the free distance to constraint length ratio of the ldpc convolutional codes exceeds the minimum distance to block length ratio of corresponding ldpc block codes .

type ia supernovae ( sne ) are believed to be the thermonuclear explosion of a c+o white dwarf ( wd ) that is destabilized when its mass approaches the chandrasekhar limit by accretion of material from a binary companion . after the central regions ignite , a burning front propagates outwards , consuming the entire star and leading to a characteristic ejecta structure with @xmath0 0.7 m@xmath3 of fe - peak nuclei ( primarily in the form of @xmath2ni that powers the sn light curve ) in the inside , and an equal amount of intermediate mass elements on the outside . while the amount of @xmath2ni can vary between @xmath0 0.3 m@xmath3 and @xmath0 1 m@xmath3 , the stratification of the ejecta remains constant across sne ia subtypes @xcite . the simple structure of the ejecta results in uniform light curves and spectra and make sn ia useful distance indicators ( e.g. , * ? ? ? * ; * ? ? ? depending upon the nature of the wd companion , sn ia progenitors are divided into either single degenerate @xcite or double degenerate cases @xcite . finding clear evidence supporting one of these two scenarios has become a cornerstone in stellar astrophysics , as unknown evolutionary effects associated with the progenitors may introduce systematic trends that could limit the precision of cosmological measurements based on sn ia @xcite . most type ia supernova remnants ( snrs ) reflect the simplicity of the explosion . galactic snrs such as tycho and sn 1006 appear remarkably symmetric @xcite , unlike their core collapse counterparts , such as cas a @xcite or younger extragalactic sne such as sn 1980k or sn 1993j @xcite . the morphology and x - ray spectra of snrs with known ages and a type ia classification also indicate that the progenitors do not modify their surroundings in a strong way in particular , while absorption features seen in the spectra of type ia sne suggest small cavities of radius @xmath7 10@xmath8 cm @xcite , there is little evidence for large ( @xmath9 @xmath0 3 30 pc ) wind - blown cavities around the explosion site @xcite . however , @xcite recently showed that the candidate type ia rcw 86 is currently interacting with a cavity wall and that the cavity has a radius of @xmath0 12 pc , in the middle of the range of sizes expected from the accretion wind models @xcite . the lack of a sizeable population of ia snr inside of cavities is particularly relevant as the current paradigm for sn ia progenitors in the single degenerate channel predicts fast , optically thick outflows from the wd surface which would leave behind such cavities ( e.g. * ? ? ? additionally , the dynamical and spectral properties of young ia snrs like tycho and 050967.5 are more consistent with an interaction with a constant density interstellar medium ( ism ) @xcite . a notable exception to this is kepler s snr ( g4.5 + 6.8 ; kepler ) , shown in x - rays in figure [ fig : kepler ] . kepler s snr has recently been firmly established as the remnant of a type ia sn , based on the o / fe ratio observed in the x - ray spectrum @xcite . however , unlike many other type ia snrs , the morphology of this @xmath10 400 yr old remnant shows a strong bilateral symmetry , and optical observations reveal clear signs of an interaction between the blastwave and a dense , nitrogen enhanced circumstellar shell with a mass of @xmath0 1m@xmath3 @xcite . near infrared observations also show evidence for circumstellar knots akin to the quasi - stationary floculi of cassiopeia a @xcite . this suggests that either the progenitor of kepler s sn or its binary companion might have been relatively massive , creating a bowshock shaped circumstellar medium ( csm ) structure as the system lost mass and moved against the surrounding ism . @xcite found values of @xmath11 @xmath0 5@xmath12 10@xmath13 m@xmath3 yr@xmath14 and v@xmath15 = 15 km s@xmath14 are required in order to explain the morphology of kepler s bowshock . while it is likely that not all progenitor scenarios for sne ia are the same , the dense csm wind required to create the observed morphology appears to be in conflict with observations of the recent nearby sn ia 2011fe in m101 @xcite . in the case of 2011fe , the radio and x - ray observations place constraints on the circumstellar environment with @xmath16 @xmath17 6 @xmath18 , or in the case of a stellar wind , @xmath11 @xmath17 6@xmath1210@xmath19 m@xmath3 yr@xmath14 @xcite . it is worth noting , however , that the recent observations of sn 2011fe probe spatial scales of @xmath0 10@xmath20 cm , corresponding to the last @xmath0 1001000 yr of the progenitor s evolution , while the blastwave of kepler s snr is probing the mass - loss history 10@xmath21 10@xmath22 yr ago . located well above the galactic plane , the snr has an angular size of 2@xmath1 in radius , but the distance is poorly constrained . @xcite used absorption to estimate a distance of 4.8 @xmath4 d@xmath23 @xmath4 6.4 kpc , while @xcite estimate that the snr is somewhat closer , with a lower limit of 3.0 kpc on the distance , based on the width of the h@xmath24 line and proper motion measurements . the non - detection of tev @xmath25-rays by hess ( f@xmath26 @xmath4 8.6 @xmath12 10@xmath27 erg @xmath28 s@xmath14 ) places a lower limit on the distance to kepler of @xmath0 6 kpc @xcite assuming a model for the gamma - ray emission @xcite , for a normal ia sn . to match the angular size , the distance would have to be even greater than 6 kpc ( @xmath7 7 kpc ) if the sn were more energetic than a normal ia , or else the blastwave would have to expand into a csm with considerable density . @xcite measured the velocity of the snr forward shock and estimated @xmath29 = 4200@xmath30 km s@xmath14 and concluded that the distance must be @xmath7 6 kpc or otherwise the sn was subenergetic . the distance estimates are incompatible with one another : either 3.0 @xmath4 d@xmath23 @xmath4 6.4 kpc , or d@xmath23 @xmath5 7 kpc , based on the non - detection of tev gamma - rays , an important new constraint . given the uncertainty on the distance , the blastwave has a radius between 1.75 @xmath31 r@xmath32 @xmath31 3.7 pc , or r@xmath32 @xmath5 4.1 pc . here we present hydrodynamical simulations of type ia sn ejecta interacting with an external medium shaped by a circumstellar wind . combining models for the pre - supernova mass - loss with models for the ejecta density distribution and composition , we simultaneously fit both the morphology of the snr and the observed x - ray spectrum within the observable constraints the age and angular size of the snr along with the bulk spectral properties . since the circumstellar environment is complicated by the bowshock structure in the north , we focus our attention on the southern portion of the snr ( see fig . [ fig : kepler ] ) , in an attempt to minimize the number of free parameters in our hydrodynamical simulations . based on our joint hydrodynamical and spectral fits , we conclude that kepler was a luminous type ia sn . if the snr is expanding into a stellar wind we require a distance to kepler of @xmath5 7 kpc . if , on the other hand , the snr is expanding into a uniform medium , the distance is somewhat closer , @xmath0 5 kpc . a 750ks _ chandra _ image of kepler s snr ( pi : s. reynolds ) is shown in figure [ fig : kepler ] , with the x - ray emission from the pie - sliced region shown in figure [ fig : spec_fit ] ( left ) . as shown in fig [ fig : spec_fit ] ( left ) , the x - ray spectrum is dominated by emission from fe k and fe l , as well as emission from intermediate mass elements such as silicon , sulfur , and argon . the amount of iron emission in the spectrum is a key constraint on the explosion , as it will allow us to discriminate between type ia explosion models @xcite . besides emission from the shocked ejecta , there is also emission from circumstellar material mainly oxygen and neon , which overlaps with the l - shell iron emission @xcite , as well as nonthermal continuum which contributes to the emission above a few kev @xcite . in the subsequent discussion on the fits to the spectrum , we do not account for either the contribution from the swept up circumstellar material or nonthermal emission in the spectral models the contribution by shocked circumstellar material to the thermal x - ray emission is small and accounts for only a few percent of the total emission @xcite , while the nonthermal emission serves to raise the continuum above @xmath0 4 kev . the processes that lead to the nonthermal emission could alter the thermal spectrum @xcite , but those affects have been shown to be small , particularly in young snrs @xcite . interesting characteristics of the spectrum include : the presence or absence of emission from iron , which will constrain the explosion energetics ; the centroids of k shell emission lines , which constrains the ionization state of the shocked gas ; and flux ratios of individual lines , such as k@xmath24 to h - like ly@xmath24 , which constrain the density of the circumstellar material . the spectra from the combined 750ks observation shown in figure [ fig : kepler ] was extracted from the level 2 _ chandra _ event list using ciao 4.1 , and weighted response matrices were computed using caldb version 4.1.3 . the broadband spectrum ( fig . [ fig : spec_fit ] , left ) shows significant emission from both fe l and fe k @xcite . given the complexity of the fe l emission at @xmath0 1 kev , we focus only on those emission lines above 1.5 kev , as they are well separated from neighboring lines . to measure the emission line centroids , we modeled the spectrum as a powerlaw continuum with a series of gaussians for each emission line . the spectral fit is shown in figure [ fig : spec_fit ] ( right ) . we chose to model the continuum as a powerlaw since @xcite showed that much of the continuum emission in kepler is nonthermal in nature ( see figs . 6 and 8 of * ? ? ? * ) . for the spectral fits , we used xspec version 11.3.2ag . we used a region nearby the supernova remnant for a local background subtraction . since we are only fitting the data above @xmath0 1.5 kev , we do not include absorption in our spectral fits . @xcite found an absorbing column of n@xmath33 = 5.2 @xmath12 10@xmath34 @xmath28 by fitting featureless regions along the remnant s rim to an absorbed srcut model . fitting only the data above 1.5 kev , we were unable to constrain the column density . however in the subsequent sections where we compare our hydrodynamical and spectral models to the data , we assume a column of 5.2 @xmath12 10@xmath34 @xmath28 so that we may compare the computed x - ray emission below 1.5 kev directly to the data . the results of our centroid fits are listed in table [ tab : fits ] . the key results in table [ tab : fits ] are the line centroids , and in the subsequent discussion , we will focus on the line centroids of of si k@xmath24 , s k@xmath24 , and fe k@xmath24 . we also point out in figure [ fig : spec_fit ] ( right ) that there is little emission at 2 kev from ly@xmath24 . this is confirmed in our spectral analysis where we are only able to set an upper limit on the ly@xmath24 emission from shocked ejecta in the south . the lack of emission at this energy will be used to constrain the csm structure and density . in this section , we present results from a modeling effort where we have coupled the hydrodynamics to a nonequilibrium ionization ( nei ) calculation to produce joint spectral and dynamical fits to kepler s snr . we describe , in general terms in 3.1 the formation of the bowshock structure in the north of kepler and how it can be used to constrain the mass - loss parameters of the progenitor system . we also describe how the choice of ia explosion model ( e.g. the explosion energy and amount of @xmath2ni that is synthesized ) imposes a constraint on the uncertain distance to the snr . in the subsequent sections , we compare the bulk spectral properties of kepler s snr to those predicted by our spectral models . the presence of a bowshock and shell of circumstellar material in the north of kepler s snr suggests that there was an outflow from the progenitor or its companion prior to the explosion . this is because the bowshock can only form at a converging flow , in this case a wind interacting with the interstellar medium , where the ism has a net inflow velocity relative to the progenitor , due to the progenitor s systemic velocity through the ism . @xcite recently modeled the dynamics of the snr blastwave through the bowshock and concluded that the dynamics of the snr blastwave are consistent with the presence of a bowshock . we can use estimates for the location of the bowshock stagnation point as a constraint on the wind parameters . generally , the location of the stagnation point , @xmath35 , is written as ( e.g. * ? ? ? * ; * ? ? ? * ) : @xmath36 where @xmath11 is the mass - loss rate , @xmath37 is the wind speed , @xmath38 is the ambient ism density , and @xmath39 is the systemic velocity of the star . in order to explain the large distance above the galactic plane , the progenitor likely had a large systemic velocity , @xmath39 @xmath0 250 km s@xmath14 @xcite , a notion confirmed by the high proper motion and radial velocities of nitrogen - rich knots in the remnant , as well as the narrow component of the h@xmath24 line in the nonradiative shock . @xcite estimated that the undisturbed ism material had a density of 10@xmath40 10@xmath41 @xmath18 , corresponding to the hot component of the ism @xcite . equation [ eqn : stag ] can be inverted to estimate the mass - loss rate of the wind : @xmath42 given the uncertainty of greater than a factor of two on the distance , the stagnation point is located at a radius of 24 pc , similar to the 23 pc assumed by @xcite . the amount of mass in the shell is @xmath10 1 m@xmath3 . @xcite have suggested that the nitrogen rich knots in the csm , combined with the solar mass of material in the shell , argue in favor of the csm being sculpted by the outflow from an asymptotic giant branch ( agb ) star . the mass - loss rates from agb stars can vary over several orders of magnitude , but the velocity of the wind is generally 1020 km s@xmath14 @xcite . using this range in velocity combined with the range in the radius of the stagnation point yields mass - loss rates of 10@xmath13 10@xmath43 m@xmath44 yr@xmath14 . to model a csm wind , the mass - loss rate and wind velocity are all that are required , since @xmath45 = @xmath46 where the wind normalization @xmath47 @xmath48 @xmath11/(4@xmath49 @xmath37 ) . we couple this model for the circumstellar environment to models for type ia sn ejecta from @xcite . we choose to focus on models ddta ( e@xmath50 = 1.4@xmath1210@xmath51 erg ) and ddtg ( e@xmath50 = 0.9@xmath1210@xmath51 erg ) . the choice of models is designed to encompass the diversity of ia explosions @xcite . in a constant density environment , the choice of explosion model does not significantly influence the radius and velocity of the forward shock as long as the kinetic energy is @xmath0 10@xmath51 erg . in a wind environment , the difference in forward shock radius between an energetic and subenergetic model , where the ejecta is approximated as a powerlaw in velocity , is a little larger , since r@xmath52 @xmath53 e@xmath54 rather than e@xmath55 @xcite . of more interest is the resultant x - ray emission , which is influenced by the density and chemical composition of the ejecta . for instance , in the ddta model , @xmath0 1m@xmath3 of @xmath2ni is produced , compared to 0.3m@xmath3 in the ddtg model . since the @xmath2ni decay is responsible for the iron observed in the ejecta , the amount of @xmath2ni produced in the sn will profoundly affect the emitted x - ray spectrum . using the vh-1 hydrodynamics code , a numerical hydrodynamics code developed at the university of virginia by j. hawley , j. blondin , and collaborators ( e.g. , * ? ? ? * ) , we modeled the expansion of the sn ejecta into the csm wind to an age of 400 yr . we varied the wind speed between 1020 km s@xmath14 and the mass - loss rate as described above . the parameter space results in a grid of models that relate the radius of the forward shock with the wind mechanical luminosity , @xmath56 = @xmath57 , for each explosion model . in figure [ fig : bounds ] , we plot the shock radius as a function of wind luminosity for each combination of wind speed and mass - loss rate . roughly speaking @xmath58 @xmath53 @xmath59 @xmath53 v@xmath60 , for a given mass - loss rate . this is because the csm density normalization is set by the wind speed higher wind velocities transport the mass further away from the progenitor , while lower velocity winds result in more mass deposited closer to the progenitor . in the higher wind velocity models , the blastwave propagates further after 400 years since it has to move through less material . as discussed earlier , the distance to kepler is not well known , ranging from 3 6.4 kpc , or @xmath5 7 kpc . in figure [ fig : bounds ] , we also note the distance to kepler on the right hand axis , using the measured angular radius of 2@xmath1 . the upper limit of 6.4 kpc @xcite on the distance rules out _ all _ ddta models on dynamical considerations alone . however , if the distance is @xmath5 7 kpc , as expected from the tev gamma - ray non - detection , then most ddta models are allowed , while a sizeable fraction of ddtg models ( @xmath11 @xmath7 6@xmath12 10@xmath13 m@xmath3 yr@xmath14 ) are no longer favored . to compare our grid of dynamical models to the observations of kepler , we have generated synthetic x - ray spectra for the emission from shocked ejecta in our hydrodynamic simulations following the methods presented in @xcite , using updated atomic data @xcite , and including radiative and ionization losses as described in @xcite . in a constant density ambient medium , the synthetic spectra are controlled by three variables : the density of the ambient material , the snr age , and the amount of collisionless heating between electrons and ions ( @xmath61 ) at the reverse shock . for our models where @xmath45 @xmath53 @xmath6 , the x - ray emission is instead influenced by the wind normalization ( which sets the density ) . higher values of @xmath11/(4@xmath49 @xmath37 ) will result in more circumstellar material closer to the explosion . we expect that this will lead to more collisional ionization in the shocked ejecta which will be directly reflected in the x - ray spectrum . the ability of our synthetic spectra to reproduce the observations is limited by the quality of the atomic data used in the code . the data for k@xmath24 blends are reasonably complete however , and we thus focus on those lines . as shown in table [ tab : fits ] , we have measured the line centroids and fluxes of several k@xmath24 lines . these can be compared directly against the synthetic spectra from our models . we focus only on the brightest lines of silicon , sulfur , and iron , and note the following additional constraints on our models ( 1 ) : the presence of fe k emission at @xmath0 6.4 kev ; ( 2 ) : the absence of significant ly@xmath24 emission at @xmath0 2.0 kev ; ( 3 ) : the amount of fe l emission below 1 kev , relative to oxygen suggests that the emission there is in fact ejecta and not shocked circumstellar material @xcite . in figure [ fig : emission ] ( left , upper panel ) we plot the x - ray spectrum from figure [ fig : spec_fit ] ( left ) compared against a subset of our spectral models . we plot the spectra from both the ddta and ddtg models for a range of mass - loss rates , assuming wind velocities of 20 km s@xmath14 . given constraints ( 1 ) and ( 3 ) above , we can rule out the ddtg models since they do not produce enough iron to match the emission seen both at 6.4 kev and fe l emission at @xmath0 1 kev . by ruling out the ddtg models , we are also essentially ruling out the distance estimate of 3 @xmath4 d@xmath23 @xmath4 6.4 kpc , if the blastwave is expanding into a @xmath62 @xmath53 @xmath6 wind . in contrast , some ddta models do reproduce the fe k@xmath24 and fe l emission . those ddta models with with @xmath11 @xmath4 4@xmath1210@xmath13 m@xmath3 yr@xmath14 are ruled out since they produce very little emission at 6.4 kev , in conflict with the data . in figure [ fig : cent_rat ] ( left , upper panels ) , we plot the line centroids from the ddta and ddtg models for various mass - loss rates for wind speeds of 20 km s@xmath14 . the hatched region in each panel corresponds to the 90% confidence on the line centroid from the spectral fitting in table [ tab : fits ] . while we have previously ruled out the ddtg models in this wind scenario based on the absence of fe emission , we are also forced to rule out those ddta models that do produce fe k emission as the predicted line centroids for nearly all models considered are at higher energies than what is measured . finally , and for comparison with other possible csm models , we note that those ddta models shown in figure [ fig : emission ] that produce fe k emission also result in significant ly@xmath24 emission at @xmath0 2.0 kev . this emission is not observed in the spectrum ( fig . [ fig : spec_fit ] ) , and our spectral fits only resulted in an upper limit on the ly@xmath24 flux ( table [ tab : fits ] ) . this is shown quantitatively in figure [ fig : cent_rat ] ( right ) where we plot the ratio of the modeled flux from si k@xmath24 to that from ly@xmath24 , for comparison against the measured ratio . in summary , based on the results of our comparisons between the observed and modeled bulk properties of the x - ray spectrum from the shocked ejecta , we must conclude that the x - ray emission from kepler s snr is not consistent with what is expected from supernova ejecta interacting with a @xmath45 @xmath53 @xmath6 wind as proposed by @xcite . as discussed above , pure wind models do not provide a good fit to the x - ray spectrum . the reason is that the low wind velocities and moderate mass - loss rates that are required to match the snr dynamics result in a significant amount of circumstellar material close to the explosion . physically , this results in a stronger reverse shock being driven into the expanding ejecta , resulting in higher average charge states this affects both the k@xmath24 line centroid as well as the amount of ly@xmath24 emission . these are two measurable quantities which do not agree with the simple models described above . the possibility exists that the circumstellar environment is more complicated than that of a simple @xmath6 wind . recurrent novae ( e.g. , * ? ? ? * ; * ? ? ? * ) or optically thick accretion winds ( e.g. , * ? ? ? * ) from the white dwarf could carve out a cavity around the progenitor , even while the companion is losing mass to a slow and dense wind . we consider the formation of a small ( @xmath0 10@xmath8 cm ) cavity inside the dense wind . the cavity could form in a number of ways ( see * ? ? ? * for a recent review ) , and time - variable d absorption spectra indicate that cavities can have sizes @xmath63 10@xmath8 cm @xcite . a small cavity such as this will not significantly alter the dynamics of the forward shock , but it will slow the ionization of the ejecta , resulting in lower charge states and dynamically younger shocked ejecta . in the bottom panel of figure [ fig : emission ] ( left ) , we plot the synthesized x - ray emission from models where the wind has a small cavity carved out of it . additionally , we show in the bottom panel of figure [ fig : cent_rat ] ( left ) the computed line centroids for these models . in the right hand panel of figure [ fig : cent_rat ] , we show the modeled flux ratio of si k@xmath24 to ly@xmath24 compared against the upper limit on the flux ratio from the x - ray spectrum , as a function of mass - loss rate . as shown in these plots , the line centroids ( si k@xmath24 and s k@xmath24 ) now agree with the measured values ( for velocities @xmath17 20 km s@xmath14 ) . additionally , these modified models show little ly@xmath24 emission at @xmath0 2 kev , as evidenced by the low flux ratios seen in the right hand panel , particularly for mass - loss rates less than 6@xmath1210@xmath13 m@xmath3 yr@xmath14 . we can thus conclude that these models are consistent with the observed x - ray spectrum . when working within the dynamical constraints imposed by the ddta models , we require a distance greater than 7 kpc , and a wind luminosity greater than @xmath0 6@xmath1210@xmath64 erg s@xmath14 . additionally , we require iron emission , agreement between the modeled and measured k@xmath24 line centroids , and little or no ly@xmath24 emission . the high mechanical wind luminosity rules out models with low values of @xmath11 , while the lower limit on the distance rules out the high mass - loss rate models . the presence of ly@xmath24 emission in the modeled spectra rule out those models with @xmath11 @xmath7 6@xmath1210@xmath13 m@xmath3 yr@xmath14 . we are left to conclude that a wind with a mass - loss rate of @xmath0 46 @xmath1210@xmath13 m@xmath3 yr@xmath14 and wind velocities @xmath0 20 km s@xmath14 sculpted the csm , and that a small cavity likely surrounded the progenitor , possibly caused by either recurrent novae or optically thick accretion winds that preceded the supernova . finally , we consider models with a uniform ambient medium , with density @xmath45 = 15 @xmath12 10@xmath65 g @xmath18 . @xcite measured the expansion of kepler s snr in the south and compared it against models for the expansion of a snr blastwave in a constant density external medium . they found that models where @xmath66 @xmath53 @xmath67 were able to reproduce the kinematics . however , they concluded that the kinematics of the blastwave in the south point to an explosion with e@xmath50 @xmath4 5 @xmath12 10@xmath68 erg for the canonical distance of 4 kpc , assuming an ambient medium density of n@xmath69 = 1 @xmath18 . higher explosion energies work if the snr is at a distance considerably greater than 4 kpc , or if the ambient medium density @xmath5 10 @xmath18 . in figure [ fig : emission ] ( _ right _ ) , we show the computed x - ray emission for ddta models in a constant density medium . as shown in fig . [ fig : emission ] , these models qualitatively reproduce the bulk features seen in the spectrum , and an ambient medium density of 25 @xmath12 10@xmath65 g @xmath18 seems to be able to reproduce both the line centroids as well as the line flux ratios . this is confirmed in figure [ fig : const_models ] , where we see that models with densities of @xmath4 3@xmath1210@xmath65 g @xmath18 produce the required ly@xmath24 emission and also predict line centroids in agreement with the measured values . for these models , the forward shock radii at 400 yrs range from 2.9 3.8 pc . for the measured angular radius of 2@xmath1 , this corresponds to a distance to kepler of 5 6.5 kpc , consistent with the distance estimate of @xcite . while the constant density models can explain the observed x - ray emission and also provide estimates for the distance to kepler that are in agreement with proper motion measurements of balmer filaments @xcite , it is difficult to reconcile a constant density csm with the bowshock structure in the north . to form a bowshock , a converging flow is required between an outflow from the progenitor and the bulk motion of the ambient ism ( in the rest frame of the progenitor ) . two alternative scenarios are considered : first , for either a wind that is not spherically symmetric but consists of equatorial and polar components @xcite , or other non - spherical outflows , such as in the recurrent nova rs oph @xcite , the csm could be sufficiently modified such that it is better described as a constant density environment , rather than a @xmath6 wind . alternatively , we can consider that kepler s snr exploded in an environment with a density gradient . @xcite modeled the evolution of a snr in a large density gradient . in figure 5(a ) of @xcite , the surface brightness profiles for the modeled remnant are qualitatively similar to the morphology of kepler . that is , a bowshock like structure forms in the direction of the density gradient . for rcw 86 , @xcite require a density contrast of @xmath4 2 between the north and south , in order to reproduce the observed x - ray emission at an age of 1800 yr . for kepler , a density contrast considerably higher than 2 would be required to produce the same morphology in just 400 yr . 160@xmath70 emission @xcite suggests a density gradient towards the north northwest of kepler , but they point out that much of that emission _ may _ arise from foreground or background sources . @xcite reached a similar conclusion from _ spitzer _ 160@xmath70 observations , noting a north south gradient in the emission at that wavelength , indicative of a density gradient in the circumstellar medium . if the bowshock structure is caused by a gradient in the density of the circumstellar environment , then that would suggest that the sn exploded in an environment that was closer to uniform in density , rather than one shaped by a slow and dense stellar wind interacting with the ism . if this were indeed the case for kepler , then it would be similar to the candidate type ia remnant g299.2 - 2.9 . in the case of that remnant , a signficant density gradient ( 3@xmath12 ) is required to explain the observed asymmetry in the x - ray morphology @xcite . however , that snr is estimated to be ten times older than kepler , so the spatial scale of the density gradient is correspondingly larger . in order to explain the observed dynamics and x - ray emission of kepler s supernova remnant , we have coupled models for ejecta from type ia sne to a model for a circumstellar medium shaped by a slow stellar wind . we have synthesized the x - ray emission from these models and compared it to the observed x - ray spectrum . our conclusions are : 1 . if the distance to kepler is taken to be less than 6.4 kpc , then the dynamics ( e.g. the forward shock radius ) are better explained by a subenergetic explosion model ( ddtg ) interacting with a slow and dense stellar wind . however , while the subenergetic models can explain the dynamics , the synthesized x - ray spectra do not agree with the observed x - ray emission . specifically , the subenergetic model does not produce enough iron to explain the observed emission . the presence of strong fe emission in the spectrum rules out all ddtg models . 2 . if the distance is taken to be greater than 7 kpc , then the dynamics can be explained by the energetic ia models . while the x - ray spectrum of kepler shows significant emission from fe k and fe l , there is little emission from h - like ly@xmath24 of intermediate mass elements such as silicon . the modeled line centroids in the ddta + wind models do not agree with the measured line centroids , even in those models with high mechanical luminosities ( i.e. high wind speeds ) . additionally , the pure wind models over predict the amount of ly@xmath24 emission relative to that which is observed . this rules out the ddta + wind models with csm winds that extend all the way to the explosion site . since the ddtg models are ruled out based on the presence of fe emission , and the ddta + wind models are ruled out based on the measured k - shell line centroids and si k@xmath24 to ly@xmath24 flux ratio , we consider the presence of a small , low density cavity around the progenitor prior to the sn . the low density cavity alters the ionization state of the shocked ejecta but does not significantly change the dynamics of the forward shock . we find that the presence of the cavity reduces both the amount of ly@xmath24 emission as well as the modeled line centroids to values consistent with the measured centroids . we find that the energetic ia model ( ddta ) interacting with a wind with velocity @xmath71 20 km s@xmath14 and mass - loss rate of @xmath71 6@xmath1210@xmath13 m@xmath3 yr@xmath14 is a good match both spectrally and dynamically , assuming a distance @xmath5 7 kpc . 4 . ddta models with a constant density ambient medium ( @xmath45 @xmath10 2 @xmath12 10@xmath65 g @xmath18 ) can explain the observed x - ray emission and place the snr at a distance that is consistent with previous estimates @xcite . in order to explain the bowshock structure in the north , a north - south density gradient in the interstellar medium is required . evidence for such a gradient is seen in 160@xmath70 emission maps of kepler and its vicinity @xcite . + thus , we conclude that in order to explain both the dynamics and x - ray emission of kepler in the context of a wind environment , the explosion was energetic , producing @xmath0 1m@xmath3 of @xmath2ni . the supernova occurred in a slow and dense csm wind with a small central cavity . the observed x - ray emission , which traces the explosion energetics , requires a distance @xmath7 7 kpc , in the csm wind model . alternatively , we find that the x - ray emission and dynamics of kepler can also be explained by a constant csm density model . these models place the snr at a distance of @xmath0 56 kpc , consistent with previous measurements . in the constant density scenario , a north south density gradient in the ism would be required in order to explain the snr morphology . d. j. p. acknowledges support from the _ chandra _ go program through grant go0 - 11094x as well as support from nasa contract nas8 - 03060 . d. j. p. is grateful for the hospitality of the weizmann institute for hosting him so that the groundwork for this paper could be performed . we thank alexa hart at sao for useful discussions regarding the abundances and wind luminosities of asymptotic giant branch stars . additionally , we thank paul plucinsky for discussions regarding the measurement of line centroids in acis spectra . lr @xmath24 & 2.67@xmath72 + norm ( 10@xmath40 photons @xmath28 s@xmath14 at 1 kev ) & 1.90@xmath73 + si k@xmath24 & [email protected] + ly@xmath24 & @xmath4 [email protected] + si k@xmath61 & [email protected] + s k@xmath24 & [email protected] + ar k@xmath24 & [email protected] + ca k@xmath24 & [email protected] + fe k@xmath24 & [email protected] + si k@xmath24 & 1.848@xmath75 + si k@xmath61 & 2.190@xmath76 + s k@xmath24 & 2.425@xmath77 + ar k@xmath24 & 3.077@xmath78 + ca k@xmath24 & 3.799@xmath79 + fe k@xmath24 & 6.450@xmath80 + si k@xmath24 & 25.9@xmath81 + si k@xmath61 & 62.2@xmath82 + s k@xmath24 & 35.9@xmath83 + ar k@xmath24 & 36.0@xmath84 + ca k@xmath24 & 57.9@xmath85 + fe k@xmath24 & 83.0@xmath86 + [ tab : fits ]

it is now well established that kepler s supernova remnant is the result of a type ia explosion . with an age of 407 years , and an angular diameter of @xmath0 4@xmath1 , kepler is estimated to be between 3.0 and 7.0 kpc distant . unlike other galactic type ia supernova remnants such as tycho and sn 1006 , and snr 0509 - 67.5 in the large magellanic cloud , kepler shows evidence for a strong circumstellar interaction . a bowshock structure in the north is thought to originate from the motion of a mass losing system through the interstellar medium prior to the supernova . we present results of hydrodynamical and spectral modeling aimed at constraining the circumstellar environment of the system and the amount of @xmath2ni produced in the explosion . using models that contain either 0.3m@xmath3 ( subenergetic ) or 1.0m@xmath3 ( energetic ) of @xmath2ni , we simulate the interaction between supernova ia ejecta and various circumstellar density models . based on dynamical considerations alone , we find that the subenergetic models favor a distance to the snr of @xmath4 6.4 kpc , while the model that produces 1m@xmath3 of @xmath2ni requires a distance to the snr of @xmath5 7 kpc . the x - ray spectrum is consistent with an explosion that produced @xmath0 1m@xmath3 of @xmath2ni , ruling out the subenergetic models , and suggesting that kepler s snr was a sn 1991t - like event . additionally , the x - ray spectrum rules out a pure @xmath6 wind profile expected from isotropic mass loss up to the time of the supernova . introducing a small cavity around the progenitor system results in modeled x - ray spectra that are consistent with the observed spectrum . if a wind shaped circumstellar environment is necessary to explain the dynamics and x - ray emission from the shocked ejecta in kepler s snr , then we require that the distance to the remnant be greater than 7 kpc .

since the discovery of fe - based high-@xmath0 superconductors @xcite , the symmetry and the gap structure of the superconducting ( sc ) state have been studied very intensively . it had been established experimentally that @xmath1-wave ( @xmath8 symmetry ) sc state is realized in usual fe - based superconductors . the gap structure in many optimally - doped high-@xmath0 compounds is nearly isotropic and fully - gapped @xcite , although some compounds show accidental nodal gap structure . in the phase diagram , the sc phase is realized next to the orthorhombic structure transition at @xmath9 , and the magnetic order is also realized at @xmath10 . below @xmath9 , the orbital polarization @xmath11 is realized @xcite , and sizable softening of shear modulus @xmath12 @xcite indicates the development of orbital fluctuations near the orthorhombic phase . strong spin fluctuations are also observed near the magnetic ordered phase . these observed orbital and spin quantum criticalities have been intensively studied theoretically , since they would be closely connected to the pairing mechanism . within the random - phase - approximation ( rpa ) , strong spin fluctuations develop in the multiorbital hubbard models for fe - based superconductors . therefore , spin - fluctuation - mediated @xmath1-wave state with sign reversal ( @xmath2-wave state ) is obtained by the rpa @xcite . however , the rpa fails to explain the non - magnetic structure transition at @xmath9 . also , orbital - fluctuation - mediated @xmath1-wave state without sign reversal ( @xmath13-wave state ) is realized by introducing the quadrupole interaction @xmath14 due to fe - ion oscillations @xcite . even for @xmath15 , strong orbital fluctuations are obtained by improving the rpa by including the vertex correction ( vc ) for the susceptibility that is dropped in the rpa @xcite : since spin and orbital fluctuations mutually develop in the self - consistent vc method , both @xmath16-wave and @xmath13-wave states can be obtained by solving the multiorbital hubbard model . to distinguish between these two sc gap states , various phase - sensitive experiments had been performed @xcite . for example , inelastic neutron scattering experiments had been performed to find the magnetic resonance scattering due to the sign reversal @xcite . however , observed `` resonance - like '' hump structure can be explained even if @xmath13-wave state is realized @xcite . impurity effect measurement is another significant phase - sensitive experiment since @xmath0 would be strongly suppressed by inter - pocket impurity scattering of cooper pairs if @xmath2-wave state is realized @xcite . many theoretical studies have been devoted so far @xcite . in ref . @xcite , the present authors studied the impurity effect due to local nonmagnetic impurities based on the realistic multiorbital model , and showed that @xmath0 in the @xmath2-wave state is strongly suppressed by inter - pocket impurity scattering of cooper pairs . however , effect of the possible non - locality of the impurity potential had not been studied . therefore , in this paper , we present a quantitative study of the _ non - local _ impurity effect on the @xmath2-wave state based on the realistic five - orbital model for fe - based superconductors . we introduce realistic models of fe - site substitutional impurity atoms , which contains the non - local modifications of potentials ( @xmath17 ) and hopping integrals ( @xmath18 ) around the impurity site , referring to a recent first principle study by nakamura _ et al_. @xcite . in various non - local impurity models with realistic parameters , the @xmath2-wave state is very fragile against impurities : original transition temperature at @xmath5 is destroyed when the residual resistivity is just @xmath19 $ ] , where @xmath7 is the mass - enhancement factor . this result is consistent with the previous theoretical study for the on - site impurity model in ref . thus , experimentally observed robustness of @xmath0 against impurities in various ( optimally - doped ) fe - based superconductors @xcite would indicate the realization of the @xmath13-wave state . the effect of fe - site substitution on @xmath0 had been carefully studied in la(fe,@xmath20)as(o , f ) for @xmath20=co , ni @xcite and in ba(fe,@xmath20)@xmath21as@xmath21 as well as ba(fe@xmath22co@xmath23cu@xmath24)@xmath21as@xmath21 @xcite . in these materials , @xmath0 is well scaled by the amount of carrier doping caused by @xmath20-atom substitutions , not by the impurity concentration irrespective of the large impurity potential @xcite . this fact would mean the robustness of the sc state against strong impurity scattering in these materials as stressed in ref . @xcite . in fe - based superconductors , impurity potential matrix @xmath25 in the _ band - diagonal basis _ is @xmath26-dependent , reflecting the multiorbital band structure @xcite . however , their @xmath26-dependences had been frequently neglected in previous studies for simplicity . in this `` constant @xmath27 model '' , both the intra - band and inter - band scatterings , @xmath28 and @xmath29 , are constant parameters . in this model , the @xmath2-wave state can be stable against impurities when @xmath30 , although it seems unrealistic since both hole - fermi surfaces ( h - fss ) and electron - fermi surfaces ( e - fss ) are composed of the common @xmath31-orbitals @xcite . in addition , obtained results are strongly changed in the unitary and intermediate regimes once the @xmath26-dependence of @xmath27 is taken into account ; see sec . [ sec : senga - dis ] . in this paper , we study the impurity effect based on the realistic two - dimensional five - orbital tight - binding model @xcite : @xmath32 where @xmath33 is the @xmath34 matrix given by the fourier transformation of the hopping integral @xmath35 introduced in ref . @xcite . here , @xmath36 represents the orbital indices , and @xmath37 is the spin index . the matrix elements of @xmath33 is given by the fourier transformation of the hopping integral @xmath38 . when the electron filling per fe - site is @xmath39 , there are two hole - pockets around the @xmath40 point , one hole - pocket around @xmath41 point , and two electron - pockets around @xmath42 and @xmath43 points . in addition , we introduce the following nonmagnetic and non - local impurity potential at site @xmath44 : @xmath45 where @xmath46 , @xmath47 and @xmath48 are the on - site , nearest - neighbor ( nn ) , and next - nearest - neighbor ( nnn ) impurity potential , and @xmath49 is the modulation of the nn or nnn hopping integrals between site @xmath50 and site @xmath51 . the present impurity potential model is depicted in fig . [ fig : imp - realspace ] . in the present model , the @xmath52-matrix due to infinite number of impurity scattering processes is given as @xmath53 where @xmath54 is the free green function in real space , given by the fourier transformation of @xmath55 , and @xmath56 is the fermion matsubara frequency . when the impurity concentration is dilute ( @xmath57 ) , the normal self - energy above @xmath0 is well approximated by the @xmath52-matrix approximation . it is given by @xmath58 where @xmath59 is given by the fourier transformation of eq . ( [ eqn : t - def ] ) . the retarded ( advanced ) self - energy is given by the analytic continuation @xmath60 : @xmath61 . then , the green function in the band - diagonal basis is @xmath62 , where @xmath63 is the band basis , @xmath64 is the dispersion of the @xmath63th band , and @xmath65 is the quasiparticle damping rate due to impurities . @xmath66 can be rewritten as @xmath67 where @xmath68 is the integration on the fs@xmath63 . when the impurity concentration is low enough , the inter - band contribution to the conductivity is negligible . if we drop the current vertex correction ( cvc ) , the conductivity is given by @xmath69 or @xmath70 , @xmath71 , @xmath72 is the quasiparticle velocity , and @xmath73 is the inter - layer distance . however , to obtain the exact conductivity for @xmath57 , the cvc should be taken into account . the exact expression for the conductivity is given as @xmath74 nm that corresponds to ba122 compounds . @xmath75 is the total velocity with cvc , which is given by solving the following bethe - salpeter equation : @xmath76 where @xmath77 is the @xmath52-matrix in the band - diagonal basis . here , we neglect the inter - band contribution since it is negligible when @xmath78 . next , we study the impurity effect on @xmath0 . the gap equation at @xmath79 in the band basis is given by @xmath80 where @xmath81 is the sc gap function in the band - diagonal basis , and @xmath82 is the pairing interaction . @xmath83 , and @xmath84 is the linearized anomalous self - energy is given as @xmath85 by solving the gap equation ( [ eqn : gap - eq ] ) together with eqs . ( [ eqn : nsigma ] ) and ( [ eqn : asigma ] ) , the impurity effect on @xmath0 for @xmath57 is exactly calculated within the bcs theory . the reduction in @xmath0 is caused by the function @xmath86 , which vanishes if @xmath81 is independent of band and momentum ( = isotropic @xmath13-wave state ) according to eqs . ( [ eqn : gamma ] ) and ( [ eqn : asigma ] ) . then , the independence of @xmath0 against impurities is derived from this relation , known as the anderson theorem . however , anderson theorem is totally violated in the @xmath2-wave state . in the present study , we consider the case of spin fluctuation mediated intra - orbital repulsive interaction between e - fs and h - fs . then , the spin susceptibility is approximately diagonal with respect to the @xmath31-orbital , since it is strongly enhanced by the intra - orbital repulsion @xmath87 . that is , spin - spin correlation develops only in the same @xmath31 orbital . by taking this fact into account , we introduce the following repulsive pairing interaction : @xmath88 for @xmath89e - fs and @xmath90h - fs and vise versa , and @xmath91 for @xmath92e - fs or @xmath93h - fs . @xmath94 is the unitary matrix connecting between band - basis and orbital - basis , and @xmath95 , where @xmath96 is the cut - off energy of the pairing interaction . in the present study , we put @xmath97ev and @xmath98ev . in this case , a fully - gapped @xmath2-wave state with @xmath99k is realized when @xmath100 . hereafter , we set the unit of energy is ev , and the unit of temperature is k. first , we study the non - local impurity potential @xmath101 for @xmath20-atom ( @xmath20= mn , co , ni , zn , ru ) given by the first principle study by nakamura _ the derived values of the on - site impurity potential @xmath46 are @xmath102 , @xmath103 , @xmath104 , @xmath105 and @xmath106 ( the unit is ev ) for mn , co , ni , zn and ru , respectively , as shown in table [ tab : table1 ] . thus , zn atom works as the unitary impurity scattering center , whereas other 3@xmath31 atoms ( mn , co and ni ) induce intermediate impurity scattering . the off - site impurity potentials ( @xmath47 and @xmath48 ) are very small . in addition , strong modulations of nn and nnn hopping integrals ( @xmath107 and @xmath108 ) are induced around the impurity atom . especially , @xmath109 for @xmath4 atom ru takes large positive value , reflecting the larger radius of 4@xmath31-orbitals . thanks to the nakamura s model , we can present a quantitative analysis of the impurity effect on the @xmath2-wave state , by taking the non - local nature of the impurity potential appropriately . .[tab : table1 ] @xmath110 is the averaged on - site impurity potential in nakamura s model @xcite . the unit of energy is ev . @xmath111 is the change in the local electron density at the impurity site due to @xmath110 in the present model without interaction . in nakamura s model , the on - site potential is orbital - dependent , and @xmath107 and @xmath108 are included . [ cols="^,^,^,^,^,^",options="header " , ] figure [ fig : nakamura - tc - nimp ] show the obtained @xmath0 as function of the ( a ) impurity concentration @xmath112 and ( b ) residual resistivity @xmath113 for various impurity atoms in the case of @xmath114 . although the fe - site substitution induces the `` impurity potential '' and `` carrier doping '' , we neglect the latter effect by fixing the electron filling @xmath39 , in order to concentrate on the former effect . when the mass - enhancement factors @xmath115 is finite , the reduction of @xmath0 per impurity concentration , @xmath116 , is renormalized by @xmath117 , while @xmath118 is independent of @xmath117 @xcite . therefore , both @xmath119 and @xmath120 are multiplied by @xmath115 . according to fig . [ fig : nakamura - tc - nimp ] ( a ) , the critical impurity concentration for the disappearance of @xmath0 , @xmath119 , strongly depends on the impurity atoms : @xmath121 \sim 3.5z^{-1}\ [ \%]$ ] for 3@xmath31-impurities ( mn , co , ni , zn ) , while @xmath122 $ ] for ru - impurities . in contrast , the values of @xmath123 shown in fig . [ fig : nakamura - tc - nimp ] ( b ) are almost independent of impurities for @xmath3-impurity atoms ( @xmath124 $ ] ) , while @xmath125 $ ] for ru - impurity . as function of ( a ) @xmath112 and ( b ) @xmath126 using the impurity potential @xmath101 for @xmath20-atom ( @xmath20= mn , co , ni , zn , ru ) . ] the residual resistivity given by nakamura parameter @xmath101 is very small ( @xmath127 $ ] ) expect for @xmath128 . one of the reasons would be that the impurity potential given in ref . @xcite may be normalized , although the bare impurity potential is required for the present study . ( the normalization would caused by the modification of the wavefunction around the impurity site in solving the kohn - sham equation , while this process is also included in the @xmath52-matrix . ) by taking this effect into account , we study the impurity potential @xmath129 with @xmath130 . to see the effect of @xmath131 , we analyze the case of @xmath132 in fig . [ fig : nakamura - tc - nimp2 ] for instance . we show ( a ) @xmath112-dependence and ( b ) @xmath126-dependence of @xmath0 . note that similar results are obtained for @xmath133 . in this case , @xmath99k is suppressed by a few percent impurity concentration . in contrast , @xmath123 is approximately independent of @xmath131 for all impurity atoms . we stress that @xmath118 is enlarged to @xmath134 $ ] by taking the impurity - induced non - local orbital order into account , because of the enlarged impurity scattering cross section @xcite . in this case , @xmath2-wave state is strongly suppressed , as we will show in sec . [ sec : inoue ] . as function of ( a ) @xmath112 and ( b ) @xmath126 using the impurity potential @xmath135 . ] for the convenience of analysis , we hereafter study the ratio @xmath136 at @xmath137 : @xmath138 is rather independent of the impurity potential strength , and also @xmath138 is essentially independent of @xmath139 according to the abrikosov - gorkov theory . in fig . [ fig : nakamura - r - m ] , we show the obtained @xmath138 as function of @xmath131 for various impurity atoms . as recognized in ref . [ fig : nakamura - tc - nimp ] , the relation @xmath140 is satisfied for all atoms . by taking the mass - renormalization factor into account , we obtain that @xmath141 $ ] for all @xmath3-impurity atoms studied in ref . @xcite , except for the ru - impurity . ( the horizontal broken line @xmath142 $ ] is the value for the infinite on - site impurity potential ( @xmath143 ) studied in ref . @xcite . ) in contrast , experimentally observed @xmath138 in optimally - doped 1111 compounds ( @xmath144 ) and 122 compounds ( @xmath145 ) is @xmath146 $ ] @xcite . therefore , the @xmath2-wave state would be too fragile against nonmagnetic impurities to explain experimental robustness of @xmath0 against impurities . similar result was reported by ikeda _ et al _ @xcite by using the nakamura s impurity model . at @xmath147 $ ] for the impurity potential @xmath148 for various @xmath20 atoms . the horizontal broken like at @xmath142 $ ] represents the value for the infinite on - site impurity potential ( @xmath143 ) studied in ref . @xcite for @xmath114 . ] according to the first principle studies in refs . @xcite , the impurity - induced change in the electron density is strongly localized at the impurity center . this fact indicates the smallness of the non - local impurity potentials @xmath149 in eq . ( [ eqn : himp ] ) . however , to obtain useful knowledge on the impurity effect , we introduce four ideal impurity potential models shown in fig . [ fig : r - pot ] ( a ) : ( i ) @xmath46-model ( only on - site potential ) , ( ii ) @xmath47-model ( only nn potential , without @xmath46 ) , ( iii ) @xmath48-model ( only nnn potential , without @xmath46 ) , and ( iv ) @xmath150-model ( plaquette impurity potential due to as - site substitution ) . note that ( ii ) and ( iii ) are very unrealistic potentials . here , we put @xmath151 for simplicity . figure [ fig : r - pot ] ( b ) shows the ratio @xmath152 at @xmath147 $ ] for models ( i)-(iv ) as function of the impurity potential , in the case of @xmath114 . the obtained @xmath138 for the @xmath48-model is as large as that for @xmath46-model , while those for other two models are smaller when @xmath153 . especially , @xmath154 $ ] is realized for the @xmath47-model for @xmath155 . we also show @xmath156 and @xmath118 with cvc in figs . [ fig : r - pot ] ( c ) and ( d ) , respectively . compared to the @xmath46-model , @xmath156 in the @xmath47 model is comparable , while @xmath126 in the @xmath47 model is much larger . for this reason , the relation @xmath157 can be achieved . -model , @xmath47-model ( without @xmath46 ) , @xmath48-model ( without @xmath46 ) , and @xmath150-model . ( b ) obtained @xmath152 as function of the impurity potential . small @xmath138 for @xmath47-model is enlarged by introducing the on - site potential @xmath46 if @xmath158 ; see fig . [ fig : r - v1 ] . ( c ) @xmath156 at @xmath147 $ ] and ( d ) @xmath118 as function of the impurity potential . , title="fig : " ] -model , @xmath47-model ( without @xmath46 ) , @xmath48-model ( without @xmath46 ) , and @xmath150-model . ( b ) obtained @xmath152 as function of the impurity potential . small @xmath138 for @xmath47-model is enlarged by introducing the on - site potential @xmath46 if @xmath158 ; see fig . [ fig : r - v1 ] . ( c ) @xmath156 at @xmath147 $ ] and ( d ) @xmath118 as function of the impurity potential . , title="fig : " ] however , the above @xmath47-model without on - site potential @xmath46 is very unrealistic . thus , we also introduce @xmath46 to the @xmath47-model : figure [ fig : r - v1 ] shows the obtained ( a ) @xmath138 and ( b ) @xmath118 as function of @xmath47 with finite @xmath46 . in the case of @xmath159 , @xmath160 quickly approaches to the value for the @xmath46-model for @xmath161 . in real compound , the relation @xmath162 is expected , as we will discuss in sec . [ sec : v1-dis ] . for the @xmath47-model with on - site potential @xmath46 . for finite @xmath163 , @xmath164 $ ] is realized for @xmath165 . ( b ) the residual resistivity @xmath118 . the value @xmath166 $ ] is shown by broken line . ] we also study the effect of hopping integral inhomogeneity around the impurity site . for a systematic study , we define @xmath167 in eq . ( [ eqn : himp ] ) as @xmath168 for the nn sites ( @xmath169 ) , and @xmath170 for the nnn sites ( @xmath171 ) . in the case of ru - impurity , @xmath172 and @xmath173 are positive according to ref . @xcite , because of the larger radius of @xmath4-orbitals . however , we also study the case @xmath174 since this situation might be realized by irradiations , by shifting the fe - ion position outside of the feas plane . in fig . [ fig : r - x1-x2 ] , we show the obtained @xmath138 and @xmath118 for various on - site potential @xmath46 as function of @xmath172 , in the case of ( a)(b ) @xmath175 and ( c)(d ) @xmath176 . in the former case , the obtained @xmath138 is very small for @xmath177 , while it quickly increases for finite @xmath46 . in contrast , @xmath138 in the latter case is large even for @xmath177 . in both cases , the residual resistivity is very small for @xmath178 , since the magnitude of the hopping integral is increased locally . therefore , the @xmath2-wave state is strongly suppressed by the hopping integral inhomogeneity in both @xmath175 and @xmath176 cases , except for a special case @xmath179 . and @xmath118 as function of @xmath180 , in the case of ( a)(b ) @xmath175 and ( c)(d ) @xmath176 . @xmath164 $ ] is realized for @xmath159 . ] to study the effect of a very wide - range impurity potential , we analyze the effect of the impurity - induced non - local orbital order ( nl - oo ) derived in ref . @xcite on the @xmath2-wave state . when this impurity - induced nl - oo is formed , the residual resistivity per 1% impurity atoms increases to @xmath181 $ ] , because of the enlarged impurity scattering cross section @xcite . this fact would resolve the problem that the residual resistivity derived from nakamura parameter is very small expect for zn - impurity atom . recent discovery of `` electronic nematic transition '' in the tetragonal phase , free from any lattice deformation , has been attracting great attention . for example , in `` detwinned '' ba(fe@xmath182co@xmath23)@xmath21as@xmath21 @xcite under very small uniaxial pressure ( @xmath183mpa ) , sizable in - plane anisotropy of resistivity emerges at @xmath184 , which is about 10k@xmath185100k higher than @xmath9 . the nematic order is also observed in bafe@xmath21(as , p)@xmath21 by the magnetic torque measurement @xcite . in ref . @xcite , the authors discussed the impurity - induced electronic nematic phase using the mean - field approximation in real space , by introducing the quadrupole interaction @xmath14 . when orbital fluctuations develop , local impurity potential induces nl - oo with @xmath186 symmetry , actually reported by stm / sts autocorrelation analyses @xcite . the large cross section of the nl - oo gives giant residual resistivity , when @xmath186 nanostructures are aligned along @xmath187-axis , the in - plane anisotropy of resistivity reaches @xmath188% , consistently with experiments @xcite . of the @xmath2-wave state in the presence of impurity - induced nl - oo obtained for @xmath189 given in ref . ] in fig . [ fig : inoue ] , we show the obtained @xmath0 as function of @xmath112 for the ( i ) on - site impurity potential @xmath189 and ( ii ) impurity - induced @xmath186 orbital order obtained by the mean - field approximation in real space for @xmath189 and @xmath190 , shown in fig . 1 ( c ) of ref . @xcite . in the case ( ii ) , the critical impurity concentration is just @xmath191 $ ] , which is about one - third of @xmath119 for the case ( i ) . the corresponding in - plane averaged resistivity is @xmath192 $ ] , which is about four times @xmath123 for the case ( i ) . that is , @xmath123 is enlarged while @xmath119 is reduced when the impurity - induced nl - oo is realized . note that the enlarged @xmath123 is still one order of magnitude smaller than experimental values ( @xmath193 $ ] ) . therefore , impurity - induced @xmath186 orbital order should cause strong suppression of the @xmath2-wave sc state . similar behaviors ( larger @xmath123 and smaller @xmath119 ) are also realized by impurity - induced short - range af order in nearly af metals , such as under - doped cuprates @xcite . we consider that this impurity - induced nl - oo will also suppress the @xmath13-wave state moderately , because of the suppression in the density - of - states ( pseudo - gap formation ) due to the orbital ( short - range ) order . under @xmath0 , the impurity - induced nl - oo will cause the `` swiss cheese hole state '' in the @xmath13-wave state . we have introduced various non - local impurity potentials , and studied the impurity effect on the @xmath2 wave state driven by the pairing interaction in eq . ( [ eqn : v ] ) . we find the relation @xmath196 $ ] holds for @xmath3- and @xmath4-impurity atoms , whereas @xmath197 $ ] in a special model . to understand these numerical results , we analyze the averaged @xmath52-matrix between fs@xmath63 and fs@xmath198 : @xmath199 then , the averaged ratio between inter - pocket and intra - pocket scattering amplitude would be @xmath200 . s as function of @xmath201 in the case of ( a ) with cvc and ( b ) without cvc , for @xmath3- and @xmath4-impurity atoms as well as @xmath46-model , @xmath150-model , and @xmath47-model without @xmath46 : note that the last potential is unrealistic . @xmath202 corresponds to @xmath203 . the proportional relation @xmath204 becomes worse by including the cvc , especially for the @xmath47-model without @xmath46 and @xmath150-model . ] figure [ fig : r - tmat ] ( a ) shows the values of @xmath138 in various impurity models as function of @xmath205 . the cvc for the conductivity is taken into account correctly . it is found that @xmath138 is approximately scaled by @xmath206 for various kinds of impurities . this result is naturally understood since the @xmath2-wave state is suppressed by the inter - fs scattering , whereas both inter- and intra - fs scattering contribute to @xmath207 , as understood by eqs . ( [ eqn : gamma ] ) and ( [ eqn : asigma ] ) . therefore , the following relationships would be realize @xcite : @xmath208 we will discuss this issue in more detail in sec . [ sec : t - dis ] . another important finding in fig . [ fig : r - tmat ] ( a ) is that the value of @xmath209 in the five - orbital model is approximately independent of the impurity potential strength ( in both born and unitary regimes ) , for all the impurity models studied here . this fact means that the constant @xmath27-model can be applicable only for born impurities , as we will discuss in sec . [ sec : senga - dis ] . in fig . [ fig : r - tmat ] ( a ) , the values of @xmath138 for @xmath47- and @xmath150-models apparently deviate from the @xmath210 line . this fact originates from the large contribution from the cvc , which becomes important when the impurity potential has finite scattering cross section , since the forward impurity scattering is correctly subtracted by taking the cvc into account . in fact , as shown in fig . [ fig : r - tmat ] ( b ) , the relation @xmath211 is apparently improved by neglecting the cvc . therefore , we should take the cvc into account to obtain quantitatively reliable value of @xmath138 , especially for wide - range impurity potentials . in the previous section , we calculated the impurity effect on the @xmath2-wave state for various non - local impurity models . when we use the impurity model parameters obtained by a recent first principle study by nakamura _ et al_. , the @xmath2-wave state is fragile against impurities , consistently with the previous theoretical study for the on - site impurity model by onari and kontani @xcite . this result is qualitatively unchanged for different non - local impurity models , when the parameters are realistic . here , we present more detailed discussions . in fig . [ fig : r - pot ] ( a ) , we have shown that @xmath212 for the @xmath47-impurity model is as small as @xmath213 $ ] , which is still much larger than experimentally observed relations @xmath214 $ ] in single crystals @xcite . moreover , as shown in fig . [ fig : r - v1 ] ( a ) , @xmath138 is strongly enlarged by introducing the no - site potential @xmath46 , which should be much larger than @xmath47 in magnitude in real impurities . by introducing @xmath47 the electron number at the nn site of the impurity center is changed by @xmath215 for @xmath216 , according to the analysis of the present five - orbital tight - binding model . however , according to the first principle study @xcite , @xmath217 would be at most @xmath218 , meaning that @xmath162 due to the strong screening effect in real compounds . therefore , the effect of @xmath47 would be negligible in the study of the impurity effect in fe - based superconductors . we have shown in fig . [ fig : r - tmat ] that the relation @xmath219 holds in some non - local impurity potentials . to understand the reason , we consider a orbital - less square lattice model for simplicity . in the born approximation , the @xmath52-matrix for the @xmath47-model is given as @xmath220 where @xmath221 is the green function in real space at @xmath222 . the cosine terms in eq . ( [ eqn : t - kp ] ) originate from the non - locality of the impurity potential . for the intra hole - pocket scattering ( @xmath223 ) , ( [ eqn : t - kp ] ) gives @xmath224 . on the other hand , for the scattering between hole- and electron - pockets ( @xmath225 and @xmath226 ) , eq . ( [ eqn : t - kp ] ) gives @xmath227 . therefore , the relation @xmath219 would be possible when @xmath228 . in this paper , we analyzed the impurity effect based on the realistic five - orbital model with non - local impurity potentials . on the other hand , more simple two - band model with constant @xmath27 impurity potential has been frequently used @xcite . here , we discuss both the usefulness and limitations of the latter model , in which the impurity potential in the band - basis ( @xmath187 , @xmath229 ) is given as @xmath230 . if we assume @xmath28 and @xmath29 are constant , then we obtain @xmath231 where @xmath232 ( @xmath233 ) is inter - band ( intra - band ) @xmath52-matrix . here , @xmath234 , and @xmath235 is the density - of - states for each band . in the born regime @xmath236 , then @xmath237 holds , and therefore the @xmath16-wave state is fragile against impurities except when @xmath238 . in the unitary regime @xmath239 , the @xmath16-wave state is robust against impurities unless @xmath240 since eq . ( [ eqn : r - constanti ] ) decreases in proportion to @xmath241 . however , this result is totally changed in the five - orbital model , in which @xmath242 is momentum - dependent : once @xmath27 is @xmath26-dependent , then eq . ( [ eqn : r - constanti ] ) does not hold as proved in ref . instead , the relation @xmath243 holds for _ all the impurity models studied here even in the unitary regime _ , as recognized by the numerical analysis in fig . [ fig : r - tmat ] . therefore , the constant @xmath27 model is applicable to fe - based superconductors only for born impurities . of the @xmath2-wave state in the two - band constant @xmath27 model studied in ref . the horizontal axis is @xmath244 . @xmath245 corresponds to @xmath203 . ] figure [ fig : senga ] shows the @xmath246 in the two - band constant @xmath27 model as function of @xmath244 , in case that the @xmath2-wave state due to inter - band repulsion is realized . in this model , @xmath246 is given by eq . ( 12 ) of ref . note that @xmath247 . we see that the relationship @xmath248 holds only in the born regime @xmath249 . however , the relation @xmath250 is realized in the unitary and intermediate regimes except for @xmath251 . based on the constant @xmath27-model , ref . @xcite mentioned that the @xmath16-wave state with @xmath99k disappears for @xmath252 $ ] at @xmath253 in the intermediate regime @xmath254 , and tried to explain the experimental small impurity effect on @xmath0 based on the @xmath2-wave scenario assuming that @xmath255 . however , we obtain @xmath256 for the realistic impurity models of @xmath3-impurity atoms as shown in fig . [ fig : r - tmat ] . in addition , @xmath123 given in ref . @xcite is very overestimated for @xmath257 , since they apply the constant @xmath27-model to the intermediate regime and the cvc for the resistivity is dropped . in fact , we obtain @xmath258 $ ] for the impurity - induced nl - oo in fig . [ fig : inoue ] ; @xmath259 in this case would be the lower limit for realistic impurity models for fe - based superconductors . in this paper , we studied the effect of in - plane impurities . in the case of `` out - of - plane '' impurities , the radius of the impurity potential @xmath260 could be much longer than the lattice spacing @xmath261 @xcite . in this case , the relation @xmath262 might be realized because of the suppression of large angle scattering , and then the @xmath16-wave state would be robust against impurities . therefore , study of the out - of - plane impurities would not be useful to distinguish between @xmath13- and @xmath16-wave states . based on the realistic five - orbital model for fe - based superconductors , we have presented a systematic study of the impurity effect on the @xmath2-wave sc state . we studied impurity model parameters for @xmath3- and @xmath4-impurity atoms obtained by a recent first principle study @xcite , in addition to various non - local impurity models . the obtained values of @xmath212 for @xmath57 as function of @xmath201 are summarized in fig . [ fig : r - tmat ] . according to the abrikosov - gorkov theory , @xmath138 is essentially independent of @xmath139 . the main results are shortly summarized as follows : + + ( i ) for @xmath3-impurity atoms , @xmath263 is realized , and we obtain @xmath141 $ ] , which is comparable to @xmath264 $ ] for the on - site model with @xmath143 . also , @xmath265 $ ] for @xmath99k . + + ( ii ) for ru - impurity atoms , @xmath266 is realized , and both @xmath138 and @xmath267 are about half of those for @xmath3-impurity atoms . + + ( iii ) for impurity - induced nl - oo model , which gives a very wide - range impurity potential , we obtain @xmath268 and @xmath269 $ ] . similar result is obtained for the @xmath150-model and @xmath47-model with finite @xmath46 . + + ( iv ) the cvc is important to obtain reliable @xmath138 and @xmath123 : @xmath138 is approximately doubled by cvc for @xmath47- and @xmath150-impurity models , for example . + + ( v ) in case we use the `` constant @xmath27 impurity model '' by putting @xmath270 , we have to keep in the born regime @xmath271 , especially when @xmath272 . + the abovementioned results are essentially consistent with the previous theoretical study for the on - site impurity model by onari and kontani @xcite . experimentally , the critical value of @xmath126 for the disappearance of @xmath273k is @xmath193 $ ] , which means that the averaged mean - free - path is comparable to the lattice spacing ( @xmath274 nm ) as discussed in ref . therefore , experimentally observed robustness of @xmath0 against various kinds of impurities in optimally - doped fe - based superconductors @xcite would be consistent with the @xmath13-wave state . in this paper , we have introduced only the repulsive pairing interaction due to spin fluctuations . when both spin and orbital fluctuations strongly develop , inter - orbital attractive interaction and intra - orbital repulsive interaction coexist . this situation is naturally expected near the orthorhombic phase , and is actually reproduced by the sc - vc theory based on the hubbard model @xcite . then , the present study indicates that impurity - induced crossover form @xmath2-wave to @xmath13-wave states would be realized in some fe - based superconductors @xcite . we are grateful to m. sato , y. kobayashi , j. li , y. matsuda , t. shibauchi , and m. nakajma for valuable discussions . this study has been supported by grants - in - aid for scientific research from mext of japan . numerical calculations were partially performed using the yukawa institute computer facility . m. yi , d. h. lu , j .- h . chu , j. g. analytis , a. p. sorini , a. f. kemper , s .- k . mo , r. g. moore , m. hashimoto , w. s. lee , z. hussain , t. p. devereaux , i. r. fisher , z .- x . shen , pnas * 108 * , 6878 ( 2011 ) . fernandes , l. h. vanbebber , s. bhattacharya , p. chandra , v. keppens , d. mandrus , m.a . mcguire , b.c . sales , a.s . sefat , and j. schmalian , phys . lett . * 105 * , 157003 ( 2010 ) k. kuroki , s. onari , r. arita , h. usui , y. tanaka , h. kontani , and h. aoki , phys . * 101 * , 087004 ( 2008 ) . i. i. mazin , d.j . singh , m.d . johannes , and m.h . du : phys . * 101 * ( 2008 ) 057003 . a. d. christianson , e. a. goremychkin , r. osborn , s. rosenkranz , m. d. lumsden , c. d. malliakas , i. s. todorov , h. claus , d. y. chung , m. g. kanatzidis , r. i. bewley , and t. guidi , nature * 456 * , 930 ( 2008 ) . a. kawabata , s. c. lee , t. moyoshi , y. kobayashi , and m. sato , j. phys . * 77 * ( 2008 ) suppl . c 103704 ; m. sato , y. kobayashi , s. c. lee , h. takahashi , e. satomi , and y. miura , j. phys . . jpn . * 79 * ( 2009 ) 014710 ; s. c. lee , e. satomi , y. kobayashi , and m. sato , j. phys . * 79 * ( 2010 ) 023702 . j. li , y. f. guo , s. b. zhang , j. yuan , y. tsujimoto , x. wang , c. i. sathish , y. sun , s. yu , w. yi , k. yamaura , e. takayama - muromachiu , y. shirako , m. akaogi , and h. kontani , phys . b * 85 * , 214509 ( 2012 ) s. ideta , t. yoshida , i. nishi , a. fujimori , y. kotani , k. ono , y. nakashima , s. yamaichi , t. sasagawa , m. nakajima , k. kihou , y. tomioka , c. h. lee , a. iyo , h. eisaki , t. ito , s. uchida and r. arita , phys . lett . * 110 * , 107007 ( 2013 ) j .- h . chu , j. g. analytis , k. d. greve , p. l. mcmahon , z. islam , y. yamamoto , and i. r. fisher , science * 329 * , 824 ( 2010 ) ; j. j. ying , x. f. wang , t. wu , z. j. xiang , r. h. liu , y. j. yan , a. f. wang , m. zhang , g. j. ye , p. cheng , j. p. hu , and x. h. chen , phys . * 107 * , 067001 ( 2011 )

we present a systematic study of the impurity effect on @xmath0 in fe - based superconductors , assuming that the sign - reversal @xmath1-wave state due to inter - pocket repulsion ( @xmath2-wave state ) is realized . for this purpose , we introduce several realistic impurity models with non - local modifications of potentials and hopping integrals around the impurity site . when we use the impurity model parameters for @xmath3- and @xmath4-impurity atoms derived from the recent first principle study by nakamura _ et al_. , we find that the @xmath2-wave state is very fragile against impurities : the superconductivity without impurities @xmath5 is destroyed by introducing small residual resistivity @xmath6 $ ] ( @xmath7 being the mass - enhancement factor ) , consistently with the previous theoretical study for the on - site impurity model by onari and kontani . this result is essentially unchanged for different non - local impurity models with realistic parameters . we also discuss the effect of the impurity - induced non - local orbital order on the superconducting state .

entanglement is a key property that makes quantum information theory different from its classical counterpart @xcite . maximally entangled states , for example , are the basis of quantum state teleportation which has recently been demonstrated @xcite . under realistic conditions , however , one will only be able to generate partially entangled mixed states . it is then of interest to be able quantify the amount of entanglement in such states . some measures of entanglement for mixed states have been suggested recently @xcite . they are useful , for example , as upper bounds to entanglement purification protocols @xcite and to the quantum channel capacity of certain quantum communication channels @xcite . unfortunately entanglement measures for mixed states ( which are relevant in the presence of noise ) are usually quite hard to calculate analytically , although an analytic expression for the entanglement of formation of two spin-@xmath0 particles is now known @xcite . the general case remains unsolved . for some problems , however , it is not so important to know the exact amount of entanglement ( a quantity that is not unique anyway ) . it would be completely sufficient if one knew which state of a family of states has the most entanglement . to answer this question it would be sufficient to find a ( hopefully as simple as possible ) quantity that preserves the ordering of density operators with respect to entanglement , i.e. that for two measures @xmath1 and @xmath2 and any two density operators @xmath3 and @xmath4 we have that @xmath5 is equivalent to @xmath6 . in this paper we will compare the entanglement of formation for two spin-@xmath0 particles , for which a closed analytical form is known , with a quantity which was proposed in @xcite as a way to quantify the degree of entanglement of a mixed state . the basis of this ` measure of entanglement ' is the peres - horodecki criterion for the separability of bipartite systems @xcite . given a state of , for example , two spin-@xmath0 systems one calculates the partial transpose of the density operator . the state is separable exactly if the partial transpose is again a positive operator . if , however , one of the eigenvalues of the partial transpose is negative then the state is entangled . one can now imagine that the amount of entanglement is quantified by the modulus of this negative eigenvalue , i.e. the larger it is , the larger the entanglement of the state . it is important to check whether this measure indeed preserves the ordering of density operators as it has been used for this purpose in some publications . in section ii we will explain the entanglement of formation and subsequently we summarize some properties of the negative eigenvalue measure of entanglement . finally , we present in section iii both a numerical and an analytical comparison of the two measures of entanglement with respect to the ordering of density operators induced by them . in section iv we sum up the results of this paper . in this section we briefly describe some entanglement measures in particular the entanglement of formation and the negative eigenvalue measure of entanglement . there are not very many good measures of entanglement . one example is the relative entropy of entanglement @xcite . in fact , it gives rise to the most restrictive upper bound on the channel capacity of the depolarizing channel @xcite . unfortunately , even for two spin-@xmath0 particles no general analytical expression has been found for it so far although many special cases can be solved analytically . a second measure of entanglement and actually the first measure of entanglement that has been proposed for mixed states is the entanglement of formation @xcite . it basically describes the amount of entanglement that needs to be shared previously in order to be able to create a particular ensemble in a given state @xmath7 by local operations . mathematically , this means that we find that pure state ensemble that realizes the state @xmath7 and which has the smallest amount of entanglement , i.e. , @xmath8 where @xmath9 is a set of not necessarily orthogonal pure states . the entanglement of formation is known to be larger than the relative entropy of entanglement which proves that in general quantum state purification methods can not recover all the entanglement that has been invested in the creation of the quantum state @xcite . a nice feature of the entanglement of formation is the fact that it can be solved analytically for a system of two spin-@xmath0 particles @xcite . this allows for fast numerical studies as the cumbersome minimization eq . ( [ formation ] ) can be avoided . the entanglement of formation can be expressed in terms of the function @xmath10 where @xmath11 for a density operator @xmath7 one defines the spin flipped state @xmath12 where the @xmath13 denotes the complex conjugate in the standard computational basis @xmath14 . one then finds the entanglement of formation to be @xmath15 where the so called concurrence is defined by @xmath16 here , @xmath17 are the eigenvalues , in decreasing order , of the hermitean matrix @xmath18 . for properties of this measure of entanglement the reader should consult the literature @xcite . it is interesting to note that since @xmath19 as a function of the concurrence @xmath20 is a strictly monotonous function and maps the interval @xmath21 $ ] on @xmath21 $ ] @xmath20 can in fact also be regarded as a measure for entanglement . we will now consider the negative eigenvalue of the partial transpose of a density operator as a measure of entanglement . in the next section we will then compare it to the entanglement of formation . for two spin-@xmath0 particles ( which form the two systems @xmath22 and @xmath23 ) any disentangled state @xmath7 can be written as the convex sum of product states @xmath24 states which permit a representation of the form eq . ( [ separa ] ) are also called separable . for two spin-@xmath0 particles there is a simple criterion to decide whether a given state is separable or not @xcite . one calculates the partial transpose @xmath25 of the density operator @xmath7 in the computational basis . this means that we transpose only one subsystem , either subsystem @xmath22 or @xmath23 . if the resulting matrix is positive semidefinite then the density operator @xmath7 is separable ; otherwise it is not . therefore the partial transpose of an entangled state has a negative eigenvalue and the idea of the negative eigenvalue measure is to use the modulus of the negative eigenvalue to quantify the entanglement of the state @xmath7 . in a mathematical form this reads as @xmath26 where the @xmath27 are the eigenvalues of the partial transpose @xmath25 . however , we do not know whether this way of quantifying the entanglement constitutes a proper measure of entanglement . therefore we will investigate in the next section numerically whether the entanglement of formation and the negative eigenvalue measure are compatible from a different point of view . we would expect that any two ` good ' entanglement measures should generate the same ordering of the density operators . this means that for two entanglement measures @xmath1 and @xmath2 and any two density operators @xmath3 and @xmath4 we have that @xmath28 why do we expect this relation to be true ? if in one measure of entanglement @xmath3 contains more entanglement than @xmath4 then we would expect that a quantum state purification method would generate more singlets from an ensemble in state @xmath3 than ensemble @xmath4 . if @xmath2 is also a measure of entanglement then we would expect that @xmath4 would yield more singlets than @xmath3 . while this reasoning is not strict , it nevertheless indicates that eq . ( [ consist ] ) should be true for two ` good ' measures of entanglement . in the next two subsections we will now check whether eq . ( [ consist ] ) is satisfied for the entanglement of formation @xmath29 and the negative eigenvalue measure of entanglement @xmath30 . for some classes of density operators we can easily check analytically whether the relation eq . ( [ consist ] ) is true . for pure states it is sufficient to consider states of the form @xmath31 as it follows from the schmidt decomposition @xcite . the entanglement of formation @xmath29 then reduces to the von neumann entropy of entanglement while the negative eigenvalue measure yields @xmath32 both measures decrease monotonically with @xmath33 so that for pure states eq . ( [ consist ] ) is satisfied . it should be mentioned that for pure states @xmath34 with @xmath31 the concurrence @xmath20 is given by @xmath35 hence , also for arbitrary pure states @xmath36 the negative eigenvalue measure and the concurrence are connected by the simple equation @xmath37 werner states are defined as @xmath38 where @xmath39 is the singlet state and @xmath40 $ ] . the concurrence of werner states is given by @xmath41 for @xmath42 ( otherwise @xmath43 is a separable state ) while the entanglement of formation is found to be @xmath44 with @xmath45 . the negative eigenvalue measure simply gives @xmath46 which means that eq . ( [ connection ] ) is also valid for werner states . again , both measures decrease monotonically with @xmath47 and therefore eq . ( [ consist ] ) is satisfied . moreover , no counterexample to eq . ( [ consist ] ) can be constructed from pure states and werner states , i.e. no two density operators @xmath3 and @xmath4 can be found which violate eq . ( [ consist ] ) when @xmath3 is assumed to be a pure state and @xmath4 corresponds to a werner state . because of the monotony of @xmath19 in eq . ( [ ent1 ] ) this can already be seen when comparing the concurrences of @xmath3 and @xmath4 with the negative eigenvalue measure of the respective states : from eq . ( [ connection ] ) it then immediately follows that @xmath48 can not be fulfilled for any value of @xmath47 . one may therefore suspect that eq . ( [ consist ] ) also holds for arbitrary mixed states . to tackle the question whether eq . ( [ consist ] ) is generally satisfied and hence the two entanglement measures induce the same ordering we have employed monte carlo simulations . if eq . ( [ consist ] ) is violated at all , it is - roughly speaking - also of interest to investigate how ` badly ' it is violated after all . in the following subsection we present the results of the numerical test that we have performed . in order to test the ordering induced by the two entanglement measures under consideration we have numerically generated a million pairs of random density matrices with respect to a certain well defined distribution . in the case of two spin-@xmath0 particles , which is the case of interest in this paper , one has to generate random density operators - i.e. linear , self adjoint , positive semidefinite operators @xmath7 of finite trace @xmath49 - acting on the four dimensional hilbert space isomorphic to @xmath50 . one can represent @xmath7 by the decomposition @xmath51 with pairwise orthogonal projections , where @xmath52 and @xmath53 . this in turn corresponds to a @xmath54 matrix of the form @xmath55 where the diagonal matrix @xmath56 is defined by @xmath57 . a plausible choice for the ensemble of unitary @xmath54 matrices which can be used to construct density matrices according to eq . ( [ construction ] ) is the one with the normalized haar measure on the group of unitary matrices @xmath58 which is called the circular unitary ensemble @xcite . it has been shown in @xcite that random unitary matrices representative for the circular unitary ensemble can be obtained as follows : let us set @xmath59 where the complex @xmath54 matrices @xmath60 , @xmath61 with three real parameters @xmath62 , @xmath63 , and @xmath64 are given by @xmath65 we now take @xmath66 and @xmath67 to be independent random variables with a uniform distribution in the interval @xmath68 , and for a given random number @xmath69 distributed uniformly in the interval @xmath70 we choose @xmath71 to be @xmath72 for @xmath73 . matrices generated in this way are then random unitary matrices of the wanted type ( except for a random phase which is of no significance for our purposes ) . finally , a set of random density matrices can be obtained by appropriately choosing the diagonal matrices @xmath56 . more precisely , the random vector with entries @xmath74 should be uniformly distributed on the manifold defined by @xmath53 . according to @xcite this can be achieved , e.g. , by setting @xmath75 , @xmath76 , @xmath77 , and @xmath78 , where again , @xmath79 , @xmath80 , and @xmath81 are random numbers drawn with respect to a uniform distribution in the interval @xmath70 . to investigate whether eq . ( [ consist ] ) is satisfied we now draw random density matrices from the previously described ensemble and calculate the eigenvalues of the partial transposes and the eigenvalues of the products of the respective density operators with the spin flipped states . after discarding those density matrices which correspond to separable states we then determine values for @xmath30 and @xmath29 as described in section ii . finally , we check for pairs @xmath82 whether the sign of @xmath83 and @xmath84 is identical . fig . [ fig1 ] shows a diagram in which @xmath85 is plotted versus @xmath86 for @xmath87 pairs @xmath88 of random density matrices . although there is obviously a certain correlation between @xmath89 and @xmath90 , there are dots in the second and the fourth quadrant of the diagram which are associated with pairs of states which do not satisfy eq . ( [ consist ] ) . one can therefore conclude that in the case of two spin-@xmath0 systems this relation is not satisfied for arbitrary mixed states , and hence , the ordering induced by the entanglement of formation @xmath29 and the one induced by the negative eigenvalue measure @xmath30 is not the same . @xmath85 @xmath86 it is interesting to note that most of the randomly drawn density operators do not correspond to entangled but to separable states ; so more density matrices have to be discarded than can be kept for the investigation . in fact , from the numerical simulations we can estimate the probability @xmath91 that a created mixed state is entangled as @xmath92 , which is in complete agreement with the analytical and numerical findings presented in @xcite . from the monte carlo simulation we can also give a rather accurate estimate of how likely it is to find a pair of states which violates eq . ( [ consist ] ) , given that both states are entangled . from the relative frequency of a violation for a million pairs of density matrices we obtain for the estimate of this probability @xmath93 @xmath94 @xmath30 @xmath29 now that we know that the ordering is different it is interesting to see to what extent the values of @xmath95 and @xmath96 differ for a given state @xmath7 , or how these random states are distributed in a diagram where @xmath29 is plotted versus @xmath30 . [ fig2 ] shows such a plot . since the ordering induced by @xmath29 and @xmath30 is not the same , there is no strictly monotonous function @xmath97\rightarrow[0,1/2]$ ] such that @xmath98 for all states @xmath7 . this is also obvious from fig . [ fig2 ] . in fig . [ fig3 ] again the distribution of states is shown , but this time the negative eigenvalue measure is plotted versus the concurrence @xmath20 . we can see that most of the dots are located close to the diagonal connecting @xmath99 and @xmath100 ; this diagonal corresponds to states satisfying eq . ( [ connection ] ) . note that the numerical simulation also strongly suggests that for a state @xmath7 with a certain value of @xmath101 the upper bound for the possible values of the negative eigenvalue measure @xmath95 is given by @xmath102 , that is , that in general @xmath103 holds for any state @xmath7 . furthermore , since for pure states the ordering induced by the two entanglement measures is the same but in general it is not , it is of interest to investigate the dependence of the probability that eq . ( [ consist ] ) is violated for a pair of entangled states @xmath82 on the ` purity of those states ' with respect to a certain characterization of the purity . in fig . [ fig4 ] we show the relative number of pairs @xmath82 of states which do not satisfy eq . ( [ consist ] ) versus @xmath104 based on a million pairs of entangled states . here , @xmath105 and @xmath106 are the linear entropies @xmath107 of @xmath3 and @xmath4 , respectively , as commonly employed , e.g. , in decoherence studies @xcite . obviously , pure states correspond to a vanishing linear entropy . @xmath30 @xmath20 we observe that the fraction of pairs @xmath82 of states with @xmath108 increases monotonically with the sum of the linear entropies of the respective states . this indicates that the more mixed the two states are with respect to the arithmetic mean of their linear entropies , the larger is the probability that this pair violates eq . ( [ consist ] ) . above a certain cut off no pairs of entangles states can be found at all - the numerical data shown in fig . [ fig4 ] give an estimate for the probability density of finding a pair of entangled states with a certain value of @xmath109 ( compare also @xcite ) . it should finally be mentioned that in @xcite yet another measure of entanglement incorporating the partial transpose is proposed : for a given state @xmath7 the quantity @xmath110 is taken as a measure , where again , the @xmath27 denote the eigenvalues of the partial transpose @xmath25 of @xmath7 . while the actual value of @xmath111 for a given state is of course different from that of @xmath30 , from numerical simulations we come to the same conclusion that the ordering induced by this measure is different from the one induced by the entanglement of formation @xmath29 . we have compared the entanglement of formation with a potential measure of entanglement that is given by the negative eigenvalue of the partial transpose of the density operator of the system . in particular the ordering of density operators with respect to the amount of entanglement induced by the two measures has been compared both numerically and analytically . we have shown that the negative eigenvalue measure does not induce the same ordering as the entanglement of formation . therefore we do not expect it to be a ` good ' measure of entanglement . in particular it can not , in general , be used to determine the most entangled state for a given family of density operators as it has been used previously . the authors would like to thank peter knight , vlatko vedral , and martin wilkens for discussions and useful hints . this work was supported in part by the epsrc , the european tmr research network erbfmrxct960066 and the european tmr research network erbfmrxct960087 . plenio and v. vedral , to appear in cont . september 1998 , also available as lanl e - print quant - ph/9804075 . bennett , g. brassard , c. crepeau , r. jozsa , a. peres , and w.k . wootters , phys . rev . lett . * 70 * , 1895 ( 1993 ) . d. boschi , s. branca , f. demartini , l. hardy , and s. popescu , phys . lett . * 80 * , 1121 ( 1998 ) . d. bouwmeester , j.w . pan , k. mattle , m. eibl , h. weinfurter , and a. zeilinger , nature * 390 * , 575 ( 1997 ) . bennett , d.p . divincenzo , j.a . smolin , and w.k . wootters , phys . rev . a * 54 * , 3824 ( 1996 ) . v. vedral , m.b . plenio , m.a . rippin , and p.l . knight , phys . lett . * 78 * , 2275 ( 1997 ) . v. vedral , m.b . plenio , k. jacobs , and p.l . knight , phys . a * 56 * , 4452 ( 1997 ) . v. vedral and m.b . plenio , phys . a * 57 * , 1619 ( 1998 ) . bennett , h.j . bernstein , s. popescu , and b. schumacher , phys . a * 53 * , 2046 ( 1996 ) . bennett , g. brassard , s. popescu , b. schumacher , j.a . smolin , and w.k . wootters , phys . lett . * 76 * , 722 ( 1996 ) . n. gisin , phys . lett . * 210 * , 151 ( 1996 ) . m. horodecki , p. horodecki , r. horodecki , phys . lett . * 78 * , 574 ( 1997 ) . m. murao , m.b . plenio , s. popescu , v. vedral , and p.l . knight , phys . rev . a * 57 * , r4075 ( 1998 ) . e. rains , lanl e - print quant - ph/9707002 . wootters , phys . lett . * 80 * , 2245 ( 1998 ) . m. horodecki , p. horodecki , r. horodecki , phys . lett . a * 223 * , 1 ( 1996 ) . a. peres , phys . . lett . * 77 * , 1413 ( 1996 ) . e. schmidt , math . ann . * 63 * , 433 ( 1907 ) . for a recent analysis of measures on the set of mixed states slater , lanl e - print quant - ph/9806039 . k. yczkowski and m. ku , j. phys . a * 27 * , 4235 ( 1994 ) ; see also m. poniak , k. yczkowski , and m. ku , j. phys . a * 31 * , 1059 ( 1998 ) . k. yczkowski , p. horodecki , a. sanpera , and m. lewenstein , lanl e - print quant - ph/9804024 . zurek , s. habib , j.p . paz , phys . * 70 * , 1187 ( 1993 ) .

we compare the entanglement of formation with a measure defined as the modulus of the negative eigenvalue of the partial transpose . in particular we investigate whether both measures give the same ordering of density operators with respect to the amount of entanglement . 2

carbon nanostructures are intensively studied both experimentally and theoretically , in view of their expected important implications for the development of electronic devices , flat panel displays , nanoswitches , etc . ( see , e.g. @xcite ) . the recent synthesis of stricty two - dimensional atomic crystals of carbon ( monolayers of graphite graphene ) @xcite is promising a wealth of new phenomena and applications in technology . the observation of anomalous transport properties , and , most exciting , the discovery of substantial field effect and magnetism at room temperature allows one to envisage graphene as a reasonable replacement of nanotubes in electronic applications @xcite . in the case of isolated graphene , the electronic states near the fermi level can be described in a simple manner @xcite . by symmetry , the lower and upper bands touch at the corners of the hexagonal brillouin zone . in the vicinity of these points , the dispersion relation is isotropic and linear , and the density of states at the fermi level is strictly zero . using the tight - binding approximation for the nearest neighbour interaction in the honeycomb lattice , an effective long - wavelength description of electronic states in graphene can be written in terms of a continuum model which is based on the dirac weyl equation for massless electrons in 2 + 1-dimensional space - time with the role of speed of light @xmath0 played by fermi velocity @xmath1 @xcite . the one - particle hamiltonian operator of the model takes the form @xmath2 where the @xmath3 matrices can be chosen in the form @xcite @xmath4 @xmath5 are the pauli matrices . the one - particle electronic wave function possesses 4 components : @xmath6 where subscripts @xmath7 and @xmath8 correspond to two sublattices of the honeycomb lattice and subscripts @xmath9 and @xmath10 correspond to two inequivalent fermi points . the linear dispersion law , @xmath11 , results in the density of states @xmath12 which is the density of states of two - dimensional ultrarelativistic fermi gas ( here @xmath13 is the area of the graphene sample ) . consequently , the thermal average charge of electronic excitations in graphene , @xmath14 is evidently zero , because eq.(4 ) is even in energy ( @xmath15 is the boltzmann constant ) . topological defects in graphene are disclinations in the honeycomb lattice , resulting from the substitution of a hexagon by , say , a pentagon or a heptagon ; such a disclination rolls up the graphitic sheet into a cone . more generally , a hexagon is substituted by a polygon with @xmath16 sides , where @xmath17 is an integer which is smaller than 6 . polygons with @xmath18 ( @xmath19 ) induce locally positive ( negative ) curvature , whereas the graphitic sheet is flat away from the defect , as is the conical surface away from the apex . in the case of nanocones with @xmath18 , the value of @xmath17 counts the number of sectors of the value of @xmath20 which are removed from the graphitic sheet . if @xmath19 , then @xmath21 counts the number of such sectors which are inserted into the graphene sheet . certainly , polygonal defects with @xmath22 and @xmath23 are mathematical abstractions , as are cones with a pointlike apex . in reality , the defects are smoothed , and @xmath18 counts the number of the pentagonal defects which are tightly clustered producing a conical shape ; such nanocones were observed experimentally @xcite . theory predicts also an infinite series of the saddle - like nanocones with @xmath21 counting the number of the heptagonal defects which are clustered in their central regions . saddle - like cones serve as an element which is necessary for joining parts of carbon nanotubes of differing radii and for creating schwarzite @xcite , a structure appearing in many forms of carbon nanofoam @xcite . as it was shown by using molecular - dynamics simulations @xcite , in the case of @xmath24 , a surface with a polygonal defect is more stable than a similarly shaped surface containing a multiple number of heptagons ; a screw dislocation can be presented as the @xmath25 limit of a @xmath16-gonal defect . in the present paper we shall consider the influence of topological defects in graphene on its electronic properties at finite temperature . the effects of the variation of the bond length or the mixing of @xmath26- with @xmath27-orbitals caused by curvature of the lattice surface are neglected , and our consideration , focusing on global aspects of coordination of carbon atoms , is based on the long - wavelength continuum model originating in the tight - binding approximation for the nearest neighbour interactions . we employ a general theory of planar relativistic fermionic systems in the singular vortex background @xcite ; its version for the case of massless fermions was elaborated in @xcite . the dirac weyl hamiltonian for electronic excitations on a graphitic nanocone with a pointlike apex takes the form @xmath28\right\},\ ] ] where @xmath29 is the pseudospin , @xmath30 and @xmath31 are the polar coordinates centred at the apex of the cone with metric @xmath32 and @xmath33 the electronic wave function on a graphitic nanocone obeys the mbius - strip - type condition @xcite : @xmath34 where @xmath35 is the operator exchanging sublattices , as well as inequivalent fermi points , and commuting with hamiltonian ( 6 ) . by performing singular gauge transformation @xmath36 one gets the wave function obeying usual condition @xmath37 in the meantime , hamiltonian ( 6 ) is transformed to @xmath38\right\},\ ] ] where relation ( 9 ) is recalled . thus , a topological defect in graphene is presented by a pseudomagnetic vortex with flux @xmath39 through the apex of a cone with deficit angle @xmath40 . by performing a unitary transformation which diagonalizes @xmath41 , one gets hamiltonian in the block - diagonal form @xmath42 where @xmath43\right\ } , \,\,\ , s=\pm1.\ ] ] let us consider the kernel of resolvent ( green s function ) of hamiltonian ( 16 ) @xmath44 where @xmath45 is the set of integer numbers , and @xmath46 is a complex parameter with dimension of energy . off - diagonal radial components are expressed through the diagonal ones , while the latter satisfy the second - order differential equations . all radial components behave asymptotically at large distances as outgoing waves , i.e. as @xmath47 at @xmath48 and as @xmath49 at @xmath50 , where @xmath51 and a physical sheet for the square root is chosen as @xmath52 ( @xmath53 ) . in the cases of @xmath54 all radial components are regular at small distances , i.e. at @xmath55 and at @xmath56 . in all other cases there are radial components that are irregular at @xmath57 and @xmath58 for some values of @xmath59 . this is related to the fact that in these cases hamiltonian ( 16 ) is not essentially self - adjoint and is characterized by the nonzero deficiency index . according to the weyl von neumann theory of self - adjoint operators ( see , e.g. @xcite ) , a procedure of the self - adjoint extension is implemented yielding the condition for the irregular radial components , which depends on the set of self - adjoint extension parameters . it should be emphasized that only irregular radial components can produce a piece in the density of states , which is odd in energy . thus , if charge ( 5 ) is nonvanishing , then it depends on the set of self - adjoint extension parameters . let us consider the possibility of charge generation in the cases of @xmath60 , when hamiltonian ( 16 ) is characterized by deficiency index @xmath61 , and there is only one self - adjoint extension parameter the irregular diagonal radial components are given by expressions @xmath63+\right . \left.+\theta(r'\!-\!r)[\sin\nu_\omega j_{-f}(k r)+\cos\nu_\omega j_{f}(k r)]h^{(1)}_{-f}(k r ' ) \right\}\!,\end{aligned}\ ] ] @xmath64+\right . \left.+\theta(r'-r)[\sin\nu_\omega j_{1-f}(k r)-\cos\nu_\omega j_{-1+f}(k r)]h^{(1)}_{1-f}(k r ' ) \right\}\,,\end{aligned}\ ] ] where @xmath65 & n_d=2,\,1,\,-1,\,-2,\,-3 , \\ -2s , & n_d=-6 , \end{array}\right.\ ] ] @xmath66(6-n_d)^{-1 } , & n_d=2,\,1,\,-1,\,-2,\,-3 , \\ 1/2 , & n_d=-6 , \end{array } \right.\ ] ] @xmath67 is the bessel function of order @xmath68 , @xmath69 is the first - kind hankel function of order @xmath68 , @xmath70 parameter @xmath71 is related to @xmath62 in the following way @xmath72 @xmath73 is the scale - invariance - breaking parameter of dimension of mass . the diagonal radial components for @xmath74 are regular . computing the contribution of the irregular radial components to the functional trace of the resolvent , we get @xmath75_{\rm{irreg}}=(1-\eta)\int\limits_{0}^{\infty}dr\,r\ , \left[a_{11}^{(n_c)}(r , r)+a_{22}^{(n_c)}(r , r)\right]=\ ] ] @xmath76 where @xmath77 and @xmath78 denotes the trace over pseudospinor indices only . using the relation between the resolvent trace and the density of states , @xmath79,\ ] ] we get the expression for the piece of the density of states , which is odd in energy : @xmath80 @xmath81 where the summation over @xmath82 is performed . although quantity ( 25 ) is negligible as compared to the ideal gas contribution ( 4 ) , lacking the factor of the sample area , it provides the generation of nonzero charge on some graphitic nanocones . substituting ( 25 ) into ( 5 ) , we get the thermal average of charge , @xmath83 taking into account relation ( 24 ) , one can present thermal average ( 5 ) in the form @xmath84 where @xmath85 is the contour consisting of two collinear straight lines , @xmath86 and @xmath87 , in the complex @xmath46-plane . by deforming contour @xmath85 to encircle poles of the hyperbolic tangential function on the imaginary axis , one gets @xmath88 where @xmath89 . using ( 23 ) and summing over @xmath90 , we get @xmath91^{2f}{\rm tan}\left(\frac \theta 2 + \frac \pi 4\right)+ \left[\frac{(2m+1)\pi}{\beta mv^2}\right]^{2(1-f)}{\rm cot}\left(\frac \theta 2 + \frac \pi 4\right)\right\}^{-1}\end{aligned}\ ] ] representation ( 26 ) can be regarded as the sommerfeld watson transform of the infinite sum representation ( 29 ) . thus , the charge is zero for graphitic nanocones with @xmath92 , since in these cases one has @xmath93 , see ( 21 ) . note that one has : @xmath94 for @xmath95 , @xmath96 for @xmath97 , and @xmath98 for @xmath99 . therefore the charge , if any , for the one - pentagon ( @xmath95 ) defect is of opposite sign to that for the one - heptagon ( @xmath97 ) defect and is of the same sign to that for the three - heptagon ( @xmath99 ) defect . in the high - temperature limit @xmath100 the charge tends to zero as inverse power of temperature : @xmath101 where @xmath102 is the riemann zeta function ( see , e.g. @xcite ) . in the zero - temperature limit the charge tends to finite value @xcite : @xmath103,\ ] ] where @xmath104 thus , there is a remarkable equation relating the ground state charges which are induced by the one - pentagon , one - heptagon and three - heptagon defects : @xmath105 as it is shown in @xcite , the induced ground state charge is accumulated around a defect . this suggests that the same happens for the temperature - induced charge as well , and the local density of states is modified in essential way in the vicinity of a defect . we conclude that the theory predicts the dependence of the induced charge on the self - adjoint extension parameter in the cases of graphitic nanocones with one pentagon , one heptagon and three heptagons at the apex . this is due to the fact that the density of electronic states attains an additional piece which is odd in energy , see ( 25 ) . actually , there are three possibilities : @xmath106 , @xmath107 , and @xmath108 . the question of which of the possibilities is realized has to be answered by future experimental measurements . the work of yu.a.s . was supported by grant no . 10/07-n `` nanostructure systems , nanomaterials , nanotechnologies '' of the national academy of sciences of ukraine and grant no . 05 - 1000008 - 7865 of the intas . the work of n.d.v . was supported by the young scientist fellowship grant ( no . 05 - 109 - 5333 ) of the intas . we acknowledge the partial support of the swiss national science foundation under the scopes project no . ib7320 - 110848 . 10 saito r , dresselhaus g and dresselhaus m s 2003 _ physical properties of carbon nanotubes _ ( london : imperial college press ) novoselov k s et al 2004 _ science _ * 306 * 666 novoselov k s at al 2005 _ proc . . usa _ * 102 * 10451 . geim a k and novoselov k s 2007 _ nature materials _ * 6 * 183 katsnelson m i 2007 _ mater . today _ * 10 * issue 1&2 20 wallace p r 1947 _ phys . rev . _ * 77 * 622 divincenzo d p and mele e j 1984 _ phys . rev . * 29 * 1685 semenoff g w 1984 _ phys . rev . lett . _ * 53 * 2449 sitenko yu a and vlasii n d 2007 _ nucl . b _ * 787 * 241 krishnan a et al 1997 _ nature _ * 388 * 451 park n et al 2003 _ phys . * 91 * 237204 rode a v et al 2004 _ phys . b _ * 70 * 54407 ihara s et al 1996 _ phys b _ * 54 * 14713 lammert p e and crespi v h 2000 _ phys . rev . lett . _ * 85 * 5190 charlier j c and rignanese g m 2001 _ phys . lett . _ * 86 * 5970 osipov v a , kochetov e a and pudlak m 2003 _ jetp _ * 96 * 140 cortijo a and vozmediano m a h 2007 _ nucl b _ * 763 * 293 cortijo a and vozmediano m a h 2007 _ eur . j. special topics _ * 148 * 83 sitenko yu a 1996 _ phys . b _ * 387 * 334 sitenko yu a 1997 _ phys . atom . nucl . _ * 60 * 2102 sitenko yu a 1999 _ mod . * 701 sitenko yu a 1999 _ phys . d _ * 60 * 125017 sitenko yu a 2000 _ ann . * 282 * 167 albeverio s , gesztezy f , hoegh - krohn r and holden h 1988 _ solvable models in quantum mechanics _ ( berlin : springer verlag ) jorgenson j and lang s 1993 _ basic analysis of regularized series and products ( lecture notes in mathematics _ vol 1564 ) ( berlin : springer verlag )

electronic excitations in a graphitic monolayer ( graphene ) in the long - wavelength approximation are characterized by the linear dispersion law , representing a unique example of the really two - dimensional `` ultrarelativistic '' fermionic system which in the presence of topological defects possesses rather unusual properties . a disclination that rolls up a graphitic sheet into a nanocone is described by a pointlike pseudomagnetic vortex at the apex of the cone , and the flux of the vortex is related to the deficit angle of the conical surface . a general theory of planar relativistic fermionic systems in the singular vortex background is employed , and we derive the analytical expression for the charge which is induced at finite temperature on some graphitic nanocones .

since the discovery of 51 peg b @xcite , the radial velocity ( rv ) technique has proven indispensable for exoplanet discovery . hundreds of exoplanets have been revealed by measuring the doppler wobble of the exoplanet host star @xcite , principally at visible wavelengths . to first order , the rv method yields the period and the minimum mass ( @xmath3 ) of the orbiting planet . in order to complete the characterization of a given exoplanet , one would want to measure its radius and constrain its atmospheric constituents . traditionally , this information is accessible only if the planet transits its host star with respect to our line of sight via transmission or secondary eclipse photometry . successes with these techniques have resulted in the detections of water , carbon monoxide , and methane on the hottest transiting gas giants @xcite . these gas giants orbit their host stars in days , are known as hot jupiters , and have an occurence rate of only 1@xmath4 in the exoplanet population @xcite . broadband spectroscopic measurements of transiting hot jupiter atmospheres are rarely able to resolve molecular bands , let alone individual lines , creating degeneracies in the solutions for atmospheric molecular abundances . high - resolution infrared spectroscopy has recently provided another route to the study of exoplanet atmospheres , one applicable to transiting and non - transiting planets alike . such studies capitalize on the doppler shift between the stellar and planet lines , allowing them to determine the atmosphere compositions , true masses , and inclinations of various systems . with the cryogenic infrared echelle spectrograph ( crires ) at the vlt , @xcite provided a proof - of - concept of this technique and detected carbon monoxide in the atmosphere of the transiting exoplanet hd 209458 b consistent with previous detections using hubble space telescope data @xcite . by detecting the radial velocity variation of a planet s atmospheric lines in about six hours of observations on single nights , further work has detected the dayside and nightside thermal spectra of various transiting and non - transiting hot jupiters , reporting detections of water and carbon monoxide , as well as the presence of winds and measurements of the length of day @xcite . with harps , @xcite recently observed the reflected light spectrum of 51 peg b in a similar manner , combining 12.5 hours of data taken over seven nights when the full dayside of the planet was observable . @xcite studied the hot jupiter tau boo b using keck nirspec ( near infrared spectrometer ) , confirmed the crires measurement of the planet s keplerian orbital velocity , and detected water vapor in the atmosphere of a non - transiting exoplanet for the first time . nirspec was used to observe an exoplanet s emission spectrum over @xmath52 - 3 hours each night across multiple epochs , in order to capture snapshots of the planet s line - of - sight motion at distinct orbital phases . in combination with the many orders of data provided by nirspec s cross dispersed echelle format and the multitude of hot jupiter emission lines in the infrared , @xcite achieved sufficient signal - to - noise to reveal the orbital properties of tau boo b via the doppler shifting of water vapor lines in its atmosphere . here , we continue our keck nirspec direct detection program with a study of the emission spectrum of the hot gas giant hd 88133 b , a system that allows us to test the brightness limits of this method and develop a more robust orbital dynamics model that can be applied to eccentric systems . in section [ rvs ] we present new ( stellar ) radial velocity ( rv ) observations of hd 88133 and an updated ephemeris . in section [ methods ] we outline our nirspec observations , reduction , and telluric correction method . in section [ 2dcc ] we describe the cross correlation and maximum likelihood analyses , and present the detection of the thermal spectrum of hd 88133 b. in section [ discussion ] we discuss the implications of this result for the planet s atmosphere and for future observations . the rv measurements of hd 88133 have been made under the purview of the california planet survey ( cps ; @xcite ) with the high resolution echelle spectrometer ( hires ; @xcite ) at the w.m . keck observatory . seventeen rv measurements of hd 88133 were published in an earlier study @xcite , and here we extend that data set to 55 individual rv measurements , having a baseline of eleven years ( see table [ rvmsmts ] ) . rv data are reduced with the standard cps hires configuration and reduction pipeline @xcite . doppler shifts are recovered by comparison to an iodine absorption spectrum and a modeling procedure presented in @xcite and @xcite . processed rv data are shown in figure [ rvplot ] . lcc[t ] 13014.947812 & -21.97 & 2.03 + 13015.947488 & 23.44 & 2.06 + 13016.952546 & 20.55 & 1.91 + 13044.088461 & 21.71 & 1.63 + 13044.869410 & -24.07 & 1.46 + 13045.843414 & -31.17 & 1.42 + 13046.081308 & -19.97 & 1.52 [ rvmsmts ] m / s ] . in the course of this paper , we will show that the most likely value for the keplerian orbital velocity of hd 88133 b is 40 @xmath0 15 km / s . [ rvplot ] the rv data are fit with a markov chain monte carlo technique following @xcite . eight free parameters ( six orbital parameters : the velocity semi - amplitude @xmath6 , the period of the orbit @xmath7 , the eccentricity of the orbit @xmath8 , the argument of periastron @xmath9 , the true anomaly of the planet at a given time @xmath10 , and the arbitrary rv zero point @xmath11 ; a linear velocity trend @xmath12 ; and a stellar jitter term @xmath13 as in @xcite ) having uniform priors contribute to the model @xmath14 and are simultaneously fit to the data @xmath15 . we initiate the mcmc chains at the values published by @xcite , allowing the chains to converge quickly and avoiding degeneracies in @xmath8 and @xmath9 . the likelihood function is given by @xmath16 where @xmath17 is the instrument error and @xmath18 is the stellar jitter . the stellar jitter term is added in quadrature to the uncertainty value of each rv measurement . best - fit orbital elements indicated by this analysis , as well as other relevant system parameters , are included in table [ systemproperties ] , and the best - fit velocity curve is shown in the figure [ rvplot ] . this represents a substantial improvement to the ephemeris originally published in @xcite . we combine these values with the velocities derived by our nirspec analysis described in sections [ methods ] and [ 2dcc ] to break the @xmath3 degeneracy for hd 88133 b. llc[h ] mass , @xmath19 & 1.18 @xmath0 0.06 @xmath20 & ( 1 ) + radius , @xmath21 & 1.943 @xmath0 0.064 @xmath22 & ( 2 ) + effective temperaure , @xmath23 & 5438 @xmath0 34 k & ( 1 ) + metallicity , @xmath24 $ ] & 0.330 @xmath0 0.05 & ( 1 ) + surface gravity , @xmath25 & 3.94 @xmath0 0.11 & ( 1 ) + rotational velocity , @xmath26 & 2.2 @xmath0 0.5 & ( 1 ) + systemic velocity , @xmath27 & -3.45 @xmath0 0.119 km / s & ( 3 ) + _ k _ band magnitude , @xmath28 & 6.2 & ( 4 ) + velocity semi - amplitude , @xmath6 & 32.9@xmath29 m / s & ( 5 ) + rv zero point , @xmath11 & 3.08@xmath30 m / s & ( 5 ) + velocity trend , @xmath31 & -0.0013@xmath32 m / s / yr & ( 5 ) + stellar jitter , @xmath13 & 4.68@xmath33 & ( 5 ) + indicative mass , @xmath3 & 0.27@xmath34 & ( 5 ) + mass , @xmath35 & 1.02@xmath1 & ( 6 ) + inclination , @xmath36 & 15@xmath2 & ( 6 ) + semi - major axis , @xmath37 & 0.04691 @xmath0 0.0008 au & ( 7 ) + period , @xmath7 & 3.4148674@xmath38 & ( 5 ) + eccentricity , @xmath8 & 0.05@xmath39 & ( 5 ) + argument of periastron , @xmath9 & 7.22@xmath40 & ( 5 ) + time of periastron , @xmath41 & 2454641.984@xmath42 jd & ( 5 ) + phase uncertainty , @xmath43 & 6.34@xmath44 & ( 5 ) + [ systemproperties ] we pursue multi - epoch observations having planetary features shifted with respect to the star s spectrum and the earth s atmosphere in order to develop techniques that can eventually be used on more slowly moving planets nearer the habitable zone . for nearly synchronously rotating hot jupiters , near - infrared emission from the dayside is likely to be most readily detectable . this strategy is fundamentally different from that used at crires to date since we do not allow the planet s signal to move across several pixels on the detector during one night of observations . data were taken on six nights ( 2012 april 1 and 3 , 2013 march 10 and 29 , 2014 may 14 , 2015 april 8) in _ l _ band and three nights ( 2015 november 21 , 2015 december 1 , and 2016 april 15 ) in _ k _ band in an abba nodding pattern with the nirspec instrument at the w.m . keck observatory @xcite . with a 0.4@xmath45x24@xmath45 slit nirspec has an _ l _ band resolution of 25,000 ( 30,000 in _ k _ band ) . individual echelle orders cover from 3.4038 - 3.4565/3.2567 - 3.3069/3.1216 - 3.1698/2.997 - 3.044 @xmath46 m in the _ l _ band and from 2.3447 - 2.3813/2.2743 - 2.3096/2.2085 - 2.2422/2.1464 - 2.1788/2.0875 - 2.1188/2.0319 - 2.0619 @xmath46 m in the _ k _ band . a schematic of the planet s orbit during our observations is given in figure [ schematic ] assuming the best - fit orbital parameters from our hires rv analysis in section [ rvs ] . details of our observations are given in table [ observationtable ] . lcccccc 2012 april 1 & 16018.837 & 0.89 & 0.86 & -20.96 & 140 & 1680 + 2012 april 3 & 16020.840 & 0.48 & 0.48 & -21.66 & 140 & 2219 + 2013 march 10 & 16361.786 & 0.29 & 0.33 & -11.59 & 180 & 2472 + 2013 march 29 & 16380.726 & 0.84 & 0.80 & -19.70 & 150 & 1812 + 2014 may 14 & 16791.796 & 0.18 & 0.22 & -29.27 & 120 & 1694 + 2015 april 8 & 17120.835 & 0.50 & 0.50 & -23.01 & 160 & 2938 + 2015 november 21 & 17348.129 & 0.06 & 0.68 & 29.95 & 60 & 2701 + 2015 december 1 & 17358.117 & 0.96 & 0.62 & 29.25 & 60 & 2823 + 2016 april 15 & 17166.300 & 0.54 & 0.53 & -29.15 & 80 & 2466 + [ observationtable ] we flat field and dark subtract the data using a python pipeline la @xcite . we extract one - dimensional spectra and remove bad pixels , and calculate a fourth - order polynomial continuum by fitting the data to a model telluric spectrum after the optimal source extraction . for _ l _ band data , the wavelength solution ( described to the fourth order as @xmath47 , @xmath48 is pixel number and @xmath37 , @xmath49 , @xmath50 , and @xmath51 are free parameters ) is calculated by fitting the data to a model telluric spectrum . however , since telluric lines are generally weaker near 2 @xmath46 m , the wavelength solution for _ k _ band data is calculated by fitting the data to a combination of model telluric and stellar spectra ( given that the stellar relative velocity is well known from optical data , and is later confirmed by the cross - correlation analysis described in section [ corr ] ) . we use an adapted phoenix model for our model stellar spectrum in the _ k _ band , as described in section [ stellarmodel ] @xcite . we show one order of reduced _ l _ band spectra in the top panel of figure [ data ] . we fit an instrument profile to the data and save it so that we may apply it to our stellar and planetary models . this instrument profile is similar to the formulation given in @xcite and is parameterized as a central gaussian with four left and four right satellite gaussians , all with variable widths . often , the best - fit widths of the third and fourth left and right satellite gaussians are zero . . overplotted in orange is the stellar spectrum of hd 88133 adapted from the phoenix stellar library @xcite . ] in this work , we depart from our traditional methods of division by an atmospheric standard ( typically an a star , c.f . @xcite ) and/or line - by - line telluric correction ( modeling atmospheric abundances with the terraspec software ; bender et al . , 2012 ) in favor of a more automated and repeatable technique : principal component analysis ( pca ) . this efficient method of telluric correction was also implemented by @xcite in their reduction of crires data on hd 189733 b. pca rewrites a data set in terms of its principal components such that the variance of the data set with respect to its mean or with respect to a model is reduced . the first principal component encapsulates the most variance ; the second , the second most , etc . over the course of an observation , the telluric components should vary the most as the air mass and atmospheric abundances change , and the planet lines should remain @xmath5constant . note that we observe for only 2 - 3 hours at a time and at a lower resolution than crires , so the planet lines do not smear . for a typical keplerian orbital velocity of a hot jupiter , we would have to observe for at least four hours to smear the planet lines across the nirspec detector pixels . to remove the telluric lines and any other time - varying effects , we aim to isolate only the strongest principal components . to perform the principal component analysis , ] , we first reduce our data set in ab pairs and construct a data matrix @xmath52 having @xmath53 rows and @xmath14 columns , where @xmath53 is the number of ab pairs and @xmath14 is the number of pixels ( 1024 for individual nirspec echelle orders ) . we linearly interpolate sub - pixel shifts between nods when aligning the ab pairs on the matrix grid , and then calculate the residual matrix @xmath54 according to @xmath55 where @xmath36 is the row number , @xmath56 is the column number , @xmath57 is either the mean of @xmath58 or a telluric model , and @xmath59 is the standard deviation of the values in column @xmath56 . we guide our principal component analysis with a telluric model for @xmath57 ( rather than the mean of @xmath58 ) that uses baseline values for water vapor , http://cso.caltech.edu/tau/ ] , carbon dioxide , ftp://aftp.cmdl.noaa.gov/products/trends/co2/co2_weekly_mlo.txt ] , and methane , ftp://ftp.cmdl.noaa.gov/data/greenhouse_gases/ch4/in-situ/surface/mlo/ch4_mlo_surface-insitu_1_ccgg_dailydata.txt ] abundances . for _ band data , baseline values for ozone from a reference tropical model are also included for the orders ranging from 3.12 - 3.17 and 3.26 - 3.31 @xmath46 m . next we calculate the covariance matrix @xmath60 of our mean - normalized data such that @xmath61 a singular value decomposition of the covariance matrix is then performed to find the principal components : @xmath62 where @xmath63 contains the left singular vectors ( or the eigenvectors ) , @xmath64 is a diagonal matrix of the singular values ( or the eigenvalues ) , and @xmath65 contains the right singular vectors . the first three eigenvectors , or principal components , for a 3.12 - 3.17 @xmath46 m order taken on 2013 march 29 are shown in figure [ data ] . the first component recovers changes in the total systemic signal with air mass ; the second encapsulates changes in abundances of telluric species , resulting in adjustments to line cores and wings ; and the third describes changes to the plate scale . higher order components typically reflect instrumental fringing and other small effects . we reconstruct the time - varying portion of each ab spectrum by combining the first @xmath66 principal components of the data set , given by @xmath67 , where @xmath68 is the first @xmath66 columns of @xmath63 and @xmath69 is the @xmath66 x @xmath66 upper - left part of @xmath64 . rank-@xmath66 data , @xmath70 , can be built as @xmath71 to produce a telluric - corrected spectrum , @xmath72 , each row @xmath73 of @xmath52 is divided by its corresponding un - mean normalized row in @xmath74 : @xmath75 finally , we collapse the rows of data in @xmath76 and clip regions of substantial telluric absorption ( @xmath7775% ) . depending on the order , anywhere between 30 and 60% of the data is removed by clipping out strong features . this results in a single high signal - to - noise spectrum for each night of observations . the final telluric - corrected spectrum for 2013 march 29 is given in the final panel of figure [ data ] . the version of pca described here diverges from the approach outlined in @xcite which used pca to determine the eigenvectors making up the light curves in each spectral channel . our formulation calculates the eigenvectors comprising each observed spectrum . we also guide our principle component analysis with a telluric model . the equivalent for @xcite would have been to guide the pca with vectors for air mass , water vapor content , etc . for data taken in _ band , we find that pca can reliably correct for the earth s atmosphere for all orders of data . however , for _ k _ band data , we can not effectively remove the dense , optically - thick telluric forest of co@xmath78 lines in the order spanning from 2.03 - 2.06 @xmath46 m , and we omit this wavelength range from subsequent analysis . it is essential to determine the efficacy of pca in removing telluric signatures and further ensure that we are not removing the planet s signal as well . since @xmath599.9% of the variance is explained by the first principal component , we find that the following results are roughly consistent when different numbers of principal components are removed from the data . we calculate the percent variance removed by each component and , if we assume the planet signal is on the order of 10@xmath79 of the total signal , then for most cases we would have to remove about fifteen components to delete the planet signal . we have experimented with incrementally removing up to ten principal components as input to the subsequent analysis ; the results and conclusions presented below use data with the first five principal components removed . we run a two - dimensional cross - correlation analysis ( todcor algorithm ; @xcite ) to find the optimum shifts for the stellar and planetary spectra entwined in our telluric- and instrument - corrected data . this requires accurate model stellar and planet spectra . our synthetic stellar spectrum is a phoenix model @xcite interpolated to match the published effective temperature @xmath23 , surface gravity @xmath25 , and metallicity @xmath24 $ ] of hd 88133 listed in table [ systemproperties ] . as hd 88133 has a @xmath26 @xmath80 5 km / s , instrumental broadening dominates , and we convolve the stellar model with the instrumental profile calculated in section [ reduction ] . we have computed the high - resolution thermal emission spectrum of hd 88133 b using both the scarlet @xcite and phoenix @xcite frameworks . an example of one order of our _ l _ band planet models is shown in figure [ planet ] . both models compute the thermal structure and equilibrium chemistry of hd 88133 b given the irradiation provided by the host star . models are computed for a cloud - free atmosphere with solar elemental composition @xcite at a resolving power of r @xmath77 250k , and assume perfect heat redistribution between the day and night sides . the model spectra are subsequently convolved with the instrumental profile derived in section [ reduction ] ( figure [ planet ] ) . we find consistent results for both models despite minor differences in the molecular line lists used . the scarlet model considers the molecular opacities of @xmath81 , @xmath82 , @xmath83 , hcn , @xmath84 , and @xmath85 and @xmath86 from the high - temperature exomol database @xcite , and @xmath87 , @xmath88 , @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 , and @xmath94 from the hitran database @xcite . absorption by the alkali metals ( li , na , k , rb , and cs ) is modeled based on the line strengths provided in the vald database @xcite and @xmath95-broadening prescription provided in @xcite . collision - induced broadening from @xmath96 and @xmath97 collisions is computed following @xcite . unlike scarlet , phoenix is a forward modeling code that converges to a solution based on traditional model atmosphere constraints ( hydrostatic , chemical , radiative - convective , and local thermodynamic equilibrium ) for an assumed elemental composition . the phoenix model uses similar opacities as scarlet ( for example , most of the latest linelists from exomol and hitran ) . additional line data for metal - hydrides come from @xcite ( and references therein ) . broadening of alkali lines follows @xcite . these differences are of only minor importance for this study because scarlet and phoenix use the same water linelist and water opacity dominates the spectral features across the spectral range of our observations ( figure [ planet ] ) . based on the effective temperatures of the planet and the star , the photometric contrast @xmath98 ( defined as the ratio of the planet flux to the stellar flux ) is on the order of 10@xmath99 . this is also a rough upper bound for the spectroscopic contrast @xmath100 . since the cross - correlation analyses described in section [ corr ] are not very sensitive to contrast ratios , varying the value of @xmath100 does not change our conclusions on the radial velocity of the planet , and thus the system inclination @xcite however , the specific value of @xmath101 does affect our conclusions on the composition and structure of the planet s atmosphere . that is , the overall velocity structure of the cross - correlation surface is not much affected by @xmath100 , though the size and structure of the final maximum likelihood peak near the planet s signature will be . each order of data for each epoch is cross - correlated with the model predictions to determine the cross - correlation function ( ccf ) using the todcor algorithm @xcite . this calculation results in a two - dimensional array of correlation values for shifts in the velocities of the star and planet . in testing our models , we find that the correlation coefficient between our stellar and planet models for the two orders spanning from 2.31 - 2.38 @xmath46 m is generally an order of magnitude higher than the same correlation coefficient in any other order in the _ k _ band , and so we omit them from this study . this behavior is due to the strong correlation between stellar and planetary co at r=30,000 . hd 88133 has an effective temperature of 5438 k and its co band ( and especially the co band head at 2.295 @xmath46 m ) is extremely prominent . the _ k _ band data analysis that follows is only performed on the three orders spanning from 2.10 - 2.20 @xmath46 m . the main absorbers in this region include carbon dioxide and water vapor . for each night s observation , we combine the ccf s of each order with equal weighting to produce the maximum likelihood curves shown in figure [ sxcorr ] . @xcite showed that the cross - correlation function is proportional to the log of the likelihood . at each epoch , we can easily retrieve and confirm the stellar velocity , as shown by the single strong peak in panel a of figure [ sxcorr ] . the stellar velocity is dominated by the barycentric velocity at the time of observation and the systemic velocity of hd 88133 . we are insensitive to the reflex motion of the star , which is on the order of 0.01 km / s . the retrieval of the planet velocity @xmath102 is more complex , as evidenced by the multiple peaks in panels b - j of figure [ sxcorr ] , and requires combining the data from multiple epochs . though there are many peaks and troughs in the maximum likelihood curves produced for each night s observation ( figure [ sxcorr ] ) , only one peak per night represents the properly registered correlation of the data with the model planetary spectrum . this is not to say that most of the maximum likelihood peaks in figure [ sxcorr ] are spurious ; rather , they are the results of unintended systematic structure in the cross - correlation space . for _ l _ band data from 3.0 - 3.4 @xmath46 m taken on 2012 april 1 and 3 , 2013 march 10 and 29 , 2014 may 14 , and 2015 april 8 , respectively . ( h ) - ( j ) : maximum likelihood function of @xmath102 for _ k _ band data from 2.10 - 2.20 @xmath46 m taken on 2012 november 21 , 2015 december 1 , and 2016 april 15 , respectively . note in ( b ) - ( j ) , the blue dashed curve shows the maximum likelihood function for the phoenix model , the red curve shows the maximum likelihood function for the scarlet model , and the grey vertical dashed lines indicate the planetary velocity shift on that date given an orbital solution having @xmath103 = 40 km / s . based on @xmath43 , the error on the calculated planetary velocity shift is about 1.2 km / s . ] the orbit of hd 88133 b is slightly eccentric , and we calculate the velocity @xmath102 of the planet for a given epoch as a function of its true anomaly @xmath10 according to @xmath104 where @xmath105 is the planet s keplerian orbital velocity , @xmath9 is the longitude of periastron measured from the ascending node , @xmath8 is the eccentricity of the orbit , and @xmath11 is the combined systemic and barycentric velocities at the time of the observation . we test a range of orbital velocities from -150 to 150 km / s in steps of 1 km / s , and in turn test a range of planet masses and orbital inclinations . figure [ kp ] shows the maximum log likelihood versus the planet s keplerian orbital velocity . . note that the vertical axes can not be directly compared , and the color scheme is the same as figure [ sxcorr ] . ( a ) : normalized log likelihood curve for six nights of _ l _ band data from 3.0 - 3.4 @xmath46 m . ( b ) : normalized log likelihood curve for three nights of _ k _ band data from 2.10 - 2.20 @xmath46 m . ( c ) : normalized log likelihood for all epochs and orders of nirspec data used in panels ( a ) and ( b ) . the grey region represents the one - sigma error bars determined by jackknife sampling for data cross - correlated with the scarlet model . ( d ) : normalized log likelihood curve for six nights of _ l _ band data cross - correlated with shuffled planetary spectra . ] when the six maximum likelihood curves produced from _ l _ band data ( panels b - g in figure [ sxcorr ] ) are combined with equal weighting according to eq . [ vf ] , we see that the most likely value for the radial projection of the planet s keplerian orbital velocity @xmath103 is 41 @xmath0 16 km / s . these error bars are calculated as in @xcite by fitting a gaussian to the likelihood peak and reporting the value of @xmath106 , assuming all the points on the maximum likelihood curve are equally weighted . we deduce that the peak in the likelihood curve based on the phoenix model at @xmath560 km / s does not represent the planet s velocity , and we prove this in section [ tests ] . the same calculation applied to the three nights of _ k _ band data ( panels h - j in fig . [ sxcorr ] ) suggests that @xmath103 is 32 @xmath0 12 km / s . the combination of all nine nights of data yields @xmath103 = 40 @xmath0 15 km / s . it is this value of @xmath103 that we use to calculate the secondary velocity curve shown in the bottom panel of figure [ schematic ] . the peak near @xmath103 = 40 km / s is consistent between the two planet models cross - correlated with the data . here , given the full suite of data , we calculate error bars on the individual points with jackknife sampling . one night s worth of data is removed from the sample and the maximum likelihood calculation is repeated . the standard deviation of each point amongst the resulting eight maximum likelihood curves is proportional to the error bars on each point on the maximum likelihood curve . we note that error bars calculated by jackknife sampling and shown in figure [ kp ] are merely an estimate . in fact , for the gaussian fit , the reduced chi - squared value ( chi - squared divided by the number of degrees of freedom ) is 0.1 , indicating that the error bars are overestimated . this can be explained by the fact that there is high variance between jackknife samples , driving a high standard deviation and therefor large error bars . to examine this behavior further , we fit a gaussian distribution ( indicating the presence of a planetary signal ) and a flat line ( indicating no planetary signal ) and determine the significance of the signal . as in @xcite , we define the bayes factor @xmath107 to be the ratio of likelihoods between two models , in this case the likelihood of the gaussian distribution compared to the likelihood of the straight line . @xmath108 must be greater than 10 for a model to be very strongly preferred . for the gaussian distribution compared the the straight line , @xmath108 is nearly 430 , indicating the the gaussian approximation to the signal at 40 km / s is significantly stronger than a flat line . even if our error bars are overestimated , they are likely not overestimated by a factor of 100 . combining @xmath103 with the parameters given in table [ systemproperties ] , a @xmath103 of 40 @xmath0 15 km / s implies that the true mass of hd 88133 b is 1.02@xmath1 and its orbital inclination is 15@xmath2 . note that the values of @xmath102 implied by the most likely value of @xmath103 often , but do not always , correspond with peaks in the maximum likelihood curves for each night , as indicated by the vertical grey dashed lines in figure [ sxcorr ] . especially for nights having a small line - of - sight velocity , planetary lines may be lost in the telluric and/or stellar cross correlation residuals . we first check our detection of hd 88133 b s emission spectrum at a @xmath103 radial velocity projection of 36 km / s by varying the spectroscopic contrast @xmath100 uniformly with orbital phase . we tested nine values of @xmath100 from 10@xmath109 to 10@xmath110 , and find the maximum likelihood peak near 40 km / s is robust for @xmath100 @xmath111 10@xmath112 . we create a shuffled " planetary model by randomly rearranging chunks of each planetary atmosphere model . if the maximum likelihood peak at 36 km / s is real , then cross correlating our data with a shuffled " planetary model ( which has no coherent planet information ) should show little to no peak at the expected @xmath103 . and , indeed , the data-shuffled " planetary model cross - correlation shows no peak at 40 km / s while the peak at @xmath5 60 km / s remains for the phoenix model ( see panel d of figure [ kp ] ) . we also check our results by varying the orbital elements of the system . we obtain roughly the same values for keplerian orbital velocity ( within the error bars ) for various combinations of eccentricities down to @xmath5 0 and arguments of perihelion within 20@xmath44 of the reported value . our results are most sensitive to these orbital elements as they affect the calculations of the true anomaly and secondary velocity , and therefore the positions of the dashed vertical lines in each epoch s maximum likelihood curve shown in figure [ sxcorr ] . even with a different ephemeris , so long as the dotted vertical lines are near their current peaks , we obtain a comparable final result for the keplerian orbital velocity of the system . we note that as hd 88133 b s orbit is slightly eccentric , it is not truly tidally locked . calculations following @xcite suggest that the planet spins about 10% faster than synchronous . our strategy prefers that the planet be tidally locked so that as much of the planet s ( dayside ) emission as possible is captured when the planet has a high line - of - sight velocity relative to the star . finally , we have reprocessed the tau boo b data published by @xcite with the methods presented in section [ methods ] . the specific departure from @xcite is the use of principal component analysis to correct for telluric features in the data ( section [ pca ] ) . our analysis of five nights of _ l _ band data recovers a projected keplerian orbital velocity of 121 @xmath0 8 km / s , a mass of 5.39@xmath113 , and an orbital inclination of 50@xmath114 for tau boo b. this is in good agreement with @xcite ( @xmath103=111@xmath05 km / s ) as well as @xcite ( @[email protected] km / s ) and @xcite ( @xmath103=115@xmath011 km / s ) , thereby validating our pca - based methods . the peak in log likelihood produced solely from _ band data ( panel a of figure [ kp ] ) is an order of magnitude larger than the peak produced solely from _ k _ band data , though the values of @xmath103 preferred by each data set are consistent . though the single channel signal - to - noise ratio for our data is greater in the _ k _ band than in the _ l _ band , we only observe hd 88133 in _ k _ for three nights and only use three orders of data in our final cross - correlation analysis whereas we observe in _ l _ for six nights and use all four orders of data . furthermore , six nights of _ l _ band observations on keck yields an aggregate shot noise of @xmath5 800,000 for all epochs and all wavelength bins , suggesting that the detection of a planet having a spectroscopic contrast down to 10@xmath79 is feasible . additionally , we attempt to detect the planet s co band near 2.295 @xmath46 m , and find that effectively separating the stellar spectrum from the planet s is a complex process . future observations should consider avoiding the co band head and focusing on the co comb ( low- to moderate- angular momentum p and r branches ) itself , particularly the regions between the stellar lines where shifted planetary co should be present . these intermediate regions have @xmath115v @xmath5 60 km / s , which is certainly sufficient for the detection of shifted planetary co and especially so given the high resolution of the next generation of cross dispersed infrared echelle spectrographs , including ishell and the upgraded crires and nirspec instruments . we took the opportunity during our reprocessing of the tau boo b data to evaluate the required accuracy of the stellar model . the original @xcite analysis was performed with a stellar spectrum using a marcs solar model with adjustments made to specific line parameters . we re - run the analysis with a phoenix model for tau boo . the correct stellar velocity at each observation epoch is still recovered and the final result for the planet s @xmath103 remains unchanged . this suggests that , for the sake of detecting the planet s spectrum and thus the planet s velocity , a detailed model of the star is not necessarily required ; however a refined stellar spectrum will be critical for learning about the planet s atmosphere in detail . acquisition of both _ l _ and _ k _ band nirspec data provides a unique opportunity to constrain the atmosphere of hd 88133 b. generally , the spectroscopic contrast is dominated by water vapor in the _ l _ band and by carbon monoxide and other species in the _ k _ band . ultimately , the shape of the log likelihood peak in figure [ kp ] will provide information about the structure of the atmosphere ( e.g. , @xcite and @xcite ) and even information about the planet s rotation rate ( e.g. , @xcite and @xcite ) . we do not presently consider orbital phase variations in atmosphere dynamics , radiative transfer , and the resulting spectral signatures from the day to night sides of the hot jupiter . we plan to examine detailed properties of hd 88133 b s atmosphere in a future study . we report the detection of the emission spectrum of the non - transiting exoplanet hd 88133 b using high resolution near - infrared spectroscopy . this detection is based on the combined effect of thousands of narrow absorption lines , predominantly water vapor , in the planet s spectrum . we find that hd 88133 b has a keplerian orbital velocity of 40 @xmath0 15 km / s , a true mass of 1.02@xmath1 , and a nearly face - on orbital inclination of 15@xmath2 . direct detection of hot jupiter atmospheres via this approach is limited in that it can not measure the _ absolute _ strengths of molecular lines , relative to the photometric contrast . thus , this method will yield degeneracies between the vertical atmospheric temperature gradients and absolution molecular abundance ratios , but the relative abundances of species should be better constrained . for transiting planets having spitzer data , it should be possible to better measure absolute abundances by comparing spitzer eclipse depths and the output of our cross - correlation analyses using various planetary atmosphere models . with the further refinement of this technique and with the improved future implementation of next - generation spectrometers and coronagraphs , especially on the largest optical / infrared telescopes , we are optimistic that this method may be extended to the characterization of terrestrial atmospheres at earth - like semi - major axes . this paper shows progress in that direction by presenting an algorithm capable of removing telluric lines whilst preserving planet lines even if the planet does not move significantly during the observations .

we target the thermal emission spectrum of the non - transiting gas giant hd 88133 b with high - resolution near - infrared spectroscopy , by treating the planet and its host star as a spectroscopic binary . for sufficiently deep summed flux observations of the star and planet across multiple epochs , it is possible to resolve the signal of the hot gas giant s atmosphere compared to the brighter stellar spectrum , at a level consistent with the aggregate shot noise of the full data set . to do this , we first perform a principal component analysis to remove the contribution of the earth s atmosphere to the observed spectra . then , we use a cross - correlation analysis to tease out the spectra of the host star and hd 88133 b to determine its orbit and identify key sources of atmospheric opacity . in total , six epochs of keck nirspec _ l _ band observations and three epochs of keck nirspec _ k _ band observations of the hd 88133 system were obtained . based on an analysis of the maximum likelihood curves calculated from the multi - epoch cross correlation of the full data set with two atmospheric models , we report the direct detection of the emission spectrum of the non - transiting exoplanet hd 88133 b and measure a radial projection of the keplerian orbital velocity of 40 @xmath0 15 km / s , a true mass of 1.02@xmath1 , a nearly face - on orbital inclination of 15@xmath2 , and an atmosphere opacity structure at high dispersion dominated by water vapor . this , combined with eleven years of radial velocity measurements of the system , provides the most up - to - date ephemeris for hd 88133 .

once the higgs boson will be discovered at the lhc , it will be crucial to test its properties , and check how well they fit in the standard model ( sm ) framework . higgs boson couplings to vector bosons , heavy quarks and heavy leptons can in principle be measured by combining informations on different production and decay channels @xcite . a measurement of the higgs boson coupling to @xmath0 quarks seems presently quite challenging . on the one hand , the sm higgs production channel @xmath4 is overwhelmed by the main production process @xmath5 at the lhc @xcite . on the other hand , processes involving the @xmath6 coupling via the higgs decay @xmath7 ( for @xmath8 gev ) seem at the moment hard to manage , due to the large @xmath0 ( and , more generally , jet ) background expected from pure qcd processes . the @xmath7 decay in the higgs production via vector - boson fusion ( vbf ) has been studied in @xcite . it gives rise to four - jet final states , out of which two jets should be @xmath0-tagged . although the vbf final states have quite distinctive kinematical features ( i.e. , two forward jets with a typical transverse momentum of order @xmath9 plus a resonant @xmath0-jet pair produced centrally ) , different sources of qcd backgrounds and hadronic effects presently make the relevance of this channel for a @xmath6 coupling determination difficult to assess . for instance , triggering on @xmath10 final states must confront with the corresponding large qcd four - jet trigger rate . the @xmath11 associated production , where the higgs boson is radiated by a top - quark pair , with subsequent @xmath7 decay , could also provide a @xmath6 coupling measurement . nevertheless , the recent inclusion of more reliable qcd background estimate and detector simulation in the corresponding signal analysis @xcite , have lowered the expectations on the potential of this channel . here we report on a further process that could help in determining the @xmath12 coupling , that was recently studied in @xcite ( where more details can be found ) . we consider the higgs boson production in vbf in association with a large transverse - momentum photon ( i.e. , @xmath13 gev ) emitted centrally ( i.e. , with pseudorapidity @xmath14 ) @xmath15 where @xmath16 decays to @xmath17 , and , at the parton level , the final qcd partons are identified with the corresponding jets @xmath18 . disregarding the resonant contribution to the process coming from the @xmath19 production , the dominant feynman diagrams are the ones involving vbf ( as shown in figure 1 , where the higgs decay to @xmath17 is not shown ) . final states @xmath20 arising from photon radiation off one of the two @xmath0-quarks arising from the higgs boson decay [ via @xmath21 fall outside the experimental @xmath22 resolution window around the @xmath23 , due to the requirement of a large @xmath24 photon . -channel feynman diagrams for @xmath16 production via @xmath25.,height=264 ] adding a central photon to the @xmath26 final state , despite a further e.m . fine structure constant @xmath27 that depletes production rates , gives a number of benefits @xcite * the large ( critical ) rate for qcd multi - jet final states that characterizes the background for @xmath26 is depleted , too , by the electromagnetic coupling when requiring a further photon in the final state ; this is expected to improve triggering efficiencies of the detector ; * the large gluonic component entering the qcd background to the plain @xmath28 final state at parton level does not take part in the radiation of a large @xmath24 photon , so making part of the potential background to @xmath29 _ inactive _ ; * further dynamical _ coherence _ effects dramatically suppress the radiation of a photon in the irreducible qcd background to @xmath30 , when the photon is central ( i.e. emitted outside the typical radiation cone around the initial / final quark legs , for quarks scattered in the @xmath31-channel ) ; * a similar _ coherence _ effect depletes the @xmath32 amplitudes ( involving neutral currents ) with the respect to the @xmath3 ones ( involving charged currents ) in figure 1 , increasing the relative sensitivity to the @xmath3 coupling in the radiative channel ; then , a measurement of the @xmath30 rate could lead to a combined determination of the higgs boson couplings to @xmath0 quarks and @xmath33 vector bosons , with less contamination from the @xmath32 coupling uncertainties ; * the requirement of a central photon strongly reduces the background arising from alternative higgs boson production processes , such as the one coming from the virtual gluon fusion @xmath34 diagrams , with a photon radiated off any external quark leg . in the following , we will elaborate on a few of the previous items . in table [ tab:1 ] , the cross section for the signal and irreducible background for the process in eq . ( [ process ] ) are shown for three values of the higgs boson mass , as independently obtained by the monte carlo event generators alpgen @xcite , and madevent @xcite , with the choice of parameters described in @xcite . the following event selection , that optimizes the significance @xmath35 , has been applied @xmath36 where @xmath37 is any pair of partons in the final state , and @xmath38 , with @xmath39 the pseudorapidity and @xmath40 the azimuthal angle . for comparison , cross sections and irreducible background for the plain vbf process are also shown . llll @xmath41 & 120 gev & @xmath42 gev & @xmath43 gev + @xmath44 \;\;\;\;\;\;\;\;$ ] & 3.6 fb & @xmath452.9 fb & @xmath452.0 fb + @xmath46 $ ] & @xmath4733 fb & @xmath4838 fb & @xmath4840 fb + @xmath49 $ ] & 320 fb & @xmath45255 fb & @xmath45168 fb + @xmath50 $ ] & 103 pb & @xmath45102 pb & @xmath45 98 pb + in case the usual pattern of qed corrections held , the request of a further hard photon would keep the relative weight of signal and background unchanged with respect to the @xmath51 case . indeed , the rates for @xmath52 and its background would be related to a @xmath53 rescaling of the rates for the @xmath54 signal and its background , respectively , keeping the @xmath55 ratio approximately stable . on the other hand , both the @xmath56 signal and its background statistics would decrease according to the rescaling factor @xmath53 . consequently , if @xmath57 is the signal significance for the vbf process ( with ) without a central photon , the signal significance for @xmath52 would fall down as @xmath58 with respect to the basic vbf process . this would question the usefulness of considering the @xmath59 variant of the @xmath54 process , apart from the expected improvement in the triggering efficiency of the detectors due to the lower background rates . in table [ tab:1 ] , one can see that the qed naive expectations do not necessarily apply when restricted regions of phase space are considered ( as discussed in detail in @xcite ) . we see that the naive qed rescaling fails for the main background processes @xmath60 , whose rate drops by about a factor 3000 after requiring a central photon , due to destructive interference ( _ coherence _ ) effects discussed in @xcite . since , on the other hand , the signal cross section roughly follows the naive qed rescaling @xmath61 , the requirement of a central photon gives rise to a dramatic increase ( by more than one order of magnitude ) in the @xmath55 ratio . indeed , in table [ tab:2 ] , comparable statistical significances for the signal with and without a photon are obtained , for an integrated luminosity of @xmath62 fb@xmath63 . the impact of including a few main reducible backgrounds for @xmath64 has also been studied in @xcite , and found to be moderate . llll @xmath41 & 120 gev & @xmath42 gev & @xmath43 gev + @xmath65 & @xmath66 2.6 & @xmath66 2.0 & @xmath66 1.3 + @xmath67 & @xmath66 3.5 & @xmath66 2.8 & @xmath451.9 + apart from enhancing the @xmath55 ratio , coherence effects in @xmath68 remarkably curb the relative contribution of the @xmath69 boson fusion diagrams with respect to the @xmath70 ones ( see @xcite for further details ) . then , the @xmath71 production at the lhc can have a role not only in the determination of the @xmath72 coupling , but also for a cleaner determination of the @xmath3 coupling . the analysis presented above does not include parton - shower effects . the latter are expected to further differentiate the signal and background final - state topology and composition . a preliminary analysis of showering and central - jet veto effects points to an improvement of @xmath73 by about a factor two @xcite . the inclusion of complete showering , hadronization , and detector simulations will be needed to establish the actual potential of the process @xmath68 . i wish to thank my collaborators emidio gabrielli , fabio maltoni , mauro moretti , fulvio piccinini , and roberto pittau for the enjoyable time i had in working out with them the results discussed above . this research was partially supported by the rtn european programmes mrtn - ct-2006 - 035505 ( heptools , tools and precision calculations for physics discoveries at colliders ) , and mrtn - ct-2004 - 503369 ( quest for unification ) .

the lhc potential for a measurement of the higgs boson coupling to the @xmath0 quark in the standard model is not well established yet . we show that requiring a large transverse momentum photon in the light higgs boson production via vector - boson fusion ( with subsequent @xmath1 decay ) could provide a further handle on the @xmath2 coupling determination , and on the measurement of the @xmath3 coupling as well .

in @xcite the author constructed the universal @xmath0 link cohomology that uses dotted foams modulo local relations , and that categorifies the @xmath0 polynomial . this was done in the spirit of bar - natan s @xcite local approach to khovanov homology on one side , and khovanov s @xcite and mackaay - vaz s @xcite @xmath1 link homology on the other side ( see also @xcite for a less general construction ) . the invariant of a link ( or tangle ) is a complex of graded free @xmath2$]-modules , up to homotopy , where @xmath3 and @xmath4 are formal variables and @xmath5 is the primitive fourth root of unity . this theory corresponds to a certain frobenius algebra structure defined on @xmath6/(x^2 -hx -a),$ ] and for the case of @xmath7 it gives rise to an isomorphic version of the @xmath0 khovanov homology @xcite . the main improvement of the theory in @xcite with respect to the original khovanov homology and bar - natan s work in @xcite is the well - defined functorialily property with respect to link cobordisms relative to boundaries , with no sign indeterminacy ( for details we refer the reader to @xcite ; we remark that similar construction and results to that in @xcite were worked out by clark , morrison and walker @xcite , with no formal variables involved ) . therefore it might be worthy to have a clear understanding of the @xmath0 foam cohomology . bar - natan explained in @xcite that his extension to tangle of the khovanov homology via cobordisms modulo relations yields an improvement in computational efficiency of the khovanov homology groups . the purpose of this paper is to modify and adapt to foams the tools used in @xcite , to obtain quick @xmath0 dotted foam cohomology computations that otherwise would have taken more than a lifetime . we also give a description of a purely topological variant of this foam theory , variant in which no dots are required on cobordisms . in section [ sec : univkh ] we briefly review the main results of the work done in @xcite . we start section [ sec : fast computations ] by providing the tools needed to obtain efficient computations and explaining the algorithm . then we show that the tools we use yield an easy proof of the invariance of the foam cohomology under reidemeister moves . after that , we consider the figure eight knot and apply the algorithm to it . section [ sec : no dots ] contains the modified version of our dotted foam theory . given a tangle diagram @xmath8 with boundary @xmath9 we construct a ` formal complex ' @xmath10 $ ] that lies in the additive category @xmath11 whose objects are formally graded resolutions of @xmath8called _ webs_and whose morphisms are formal linear combinations of _ singular _ cobordisms called _ foams_whose tops and bottoms are resolutions and whose side boundaries are @xmath12 modulo certain local relations . we explain these concepts below . each crossing of @xmath8 is replaced by one of the planar pictures below : @xmath13{poscrossing.pdf } } \longrightarrow \raisebox{-13pt } { \includegraphics[height=0.4in]{orienres.pdf } } \quad \text{and } \quad \raisebox{-13pt } { \includegraphics[height=0.4in]{singres.pdf}}\ ] ] a diagram @xmath14 obtained by resolving all crossings of @xmath8 is a disjoint union of webs . a web with boundary @xmath15 is a planar graph @xmath14properly embedded in a disk @xmath16with bivalent vertices near which the two incident edges are oriented either towards the vertex or away from it , and with univalent vertices that lie on the boundary of the disk @xmath17 webs without vertices are also allowed . the @xmath0-link invariant is given by @xmath18 where the sum is over all resolutions of @xmath19 the exponents @xmath20 are determined by relations @xmath21{poscrossing.pdf } } \,= \,q\,\,\,\,\,\raisebox{-13pt } { \includegraphics[height=0.4in]{orienres.pdf } } - q^2\,\,\,\,\,\raisebox{-13pt } { \includegraphics[height=0.4in]{singres.pdf}}\hspace{2 cm } \raisebox{-13pt}{\includegraphics[height=0.4in]{negcrossing.pdf } } \,= \,q^{-1}\,\raisebox{-13pt } { \includegraphics[height=0.4in]{orienres.pdf } } - q^{-2}\ , \raisebox{-13pt } { \includegraphics[height=0.4in]{singres.pdf}}\end{aligned}\ ] ] and the _ bracket polynomial _ @xmath22 associated to a closed web @xmath14 ( its boundary @xmath15 is empty ) is evaluated via the skein relations @xmath23{unknot.pdf } } \bigcup \gamma\rangle } = ( q + q^{-1 } ) { \langle \gamma\rangle } = { \langle \raisebox{-5pt}{\includegraphics[height=0.2in]{circle2sv.pdf } } \bigcup \gamma\rangle } \\ { \langle \raisebox{-5pt}{\includegraphics[height=0.12in]{2vertweb.pdf}}\rangle } = { \langle \raisebox{-5pt}{\includegraphics[height=0.12in]{arcro.pdf}}\rangle } , \quad { \langle \raisebox{-5pt}{\includegraphics[height=0.12in]{2vertwebleft.pdf}}\rangle } = { \langle \raisebox{-5pt}{\includegraphics[height=0.12in]{arclo.pdf}}\rangle}.\end{aligned}\ ] ] a foam is an abstract cobordism between two webs @xmath24 and @xmath25 with boundary @xmath9 regarded up to boundary - preserving isotopy . we read foams as morphisms from bottom to top by convention , and we compose them by placing one cobordism on top the other . foams have _ singular arcs _ ( and/or _ singular circles _ ) where orientations disagree , and near which the facets incident with a given singular arc are compatibly oriented , inducing an orientation on that arc . specifically , the orientation of singular arcs is as in the figure below , which shows examples of _ singular saddles_. @xmath26{saddle.pdf}\ ] ] for each singular arc of a foam , there is an ordering of foam s facets that are incident with it , in the sense that one of the facets is the _ preferred facet _ for that singular arc . in some cases , we indicate facets ordering by using labels 1 and 2 , respectively , as shown above . in other cases , to prevent cluster , we draw a singular arc using a continuous red curve if the preferred facet for that singular arc is at its left in a given plane projection of a foam where the concept of ` left ' or ` right ' is given by the orientation of the arc ; otherwise , the singular arc is represented by a dashed red curve . finally , foams can have dots that are allowed to move freely along the facet they belong to , but ca nt cross singular arcs . we mod out the set of foams by the local relations @xmath27 = ( 2d , sf , s , ufo ) below . @xmath28{plane2d.pdf}}=h\raisebox{-5pt}{\includegraphics[height=0.2in]{plane1d.pdf } } + a\raisebox{-5pt}{\includegraphics[height=0.2in]{plane.pdf}},\ \ \ \ \ \ \ \raisebox{-10pt}{\includegraphics[width=0.35in]{sph.pdf } } = 0,\quad \raisebox{-10pt}{\includegraphics[width=0.35in]{sphd.pdf } } = 1 & \text{(s ) } \\ \raisebox{-20pt } { \includegraphics[height=0.6in]{surgery.pdf } } = \raisebox{-20pt } { \includegraphics[height=0.6in]{surgery1.pdf}}+ \raisebox{-20pt } { \includegraphics[height=0.6in]{surgery2.pdf}}-h \raisebox{-20pt } { \includegraphics[height=0.6in]{surgery3 } } & \text{(sf ) } \\ \raisebox{-10pt}{\includegraphics[width=0.5in]{ufolo.pdf}}=0= \raisebox{-10pt}{\includegraphics[width=0.5in]{ufolo2d.pdf}}\quad \text{and } \quad \raisebox{-10pt}{\includegraphics[width=0.5in]{ufolold.pdf}}=i= - \raisebox{-10pt}{\includegraphics[width=0.5in]{ufolord.pdf } } & \text{(ufo ) } } \ ] ] relations @xmath27 imply the following useful _ curtain identities _ : @xmath29{handy - rel1.pdf}}&= i \,\raisebox{-13pt}{\includegraphics[height=0.5in]{handy - rel3.pdf } } \hspace{2 cm } \raisebox{-13pt}{\includegraphics[height = 0.5in]{handy - rel4.pdf } } = -i \,\raisebox{-13pt}{\includegraphics [ height=0.5in]{handy - rel6.pdf } } \\ \raisebox{-18pt}{\includegraphics[height=0.6in]{curtainid.pdf } } & = -i \ , \raisebox{-18pt}{\includegraphics[height=0.6in]{curtain2sa.pdf } } \hspace{1 cm } \raisebox{-18pt}{\includegraphics[height=0.6in]{curtainid2.pdf}}= i\ , \raisebox{-18pt}{\includegraphics[height=0.6in]{curtain2sa2.pdf}}\end{aligned}\ ] ] as well as the _ cutting neck _ relation ( cn ) , where the dots belong to the preferred facets : @xmath30{cneck.pdf}}= -i \raisebox{-18pt}{\includegraphics[height=0.6in]{cneck1.pdf } } -i \raisebox{-18pt}{\includegraphics[height=0.6in]{cneck2.pdf } } + hi \raisebox{-18pt}{\includegraphics[height=0.6in]{cneck3 } } & \text{(cn ) } } \ ] ] local relations @xmath27 also give rules for _ exchanging dots _ between neighboring facets , namely : @xmath31{exch1dl.pdf } } + \raisebox{-18pt}{\includegraphics[height=.5in]{exch1dr.pdf } } = h\,\,\raisebox{-18pt}{\includegraphics[height=.5in]{exch.pdf } } \hspace{2 cm } \raisebox{-18pt}{\includegraphics[height=.5in]{exch2d.pdf } } = -a\ , \ , \raisebox{-18pt}{\includegraphics[height=.5in]{exch.pdf}}\end{aligned}\ ] ] [ lemma : isomorphisms ] local relations @xmath27 imply that the following are isomorphisms in @xmath32 @xmath33{basic - isom1.pdf } } \hspace{1.5 cm } \raisebox{-25pt}{\includegraphics[height=.8in]{basic - isom2.pdf}}\end{aligned}\ ] ] the formal complex @xmath10 $ ] is an object of the category _ kof _ : = kom(mat(@xmath11 ) ) of complexes of formal direct sums of objects in @xmath11 and its cohomology @xmath34 is a doubly graded invariant up to homotopy ( here @xmath5 is the cohomological grading and @xmath35 is the polynomial grading ) . in this section we apply to our setting bar - natan s ` divide and conquer ' approach to computations discussed in @xcite , to obtain an efficient algorithm for calculating the homology groups @xmath36 associated to a certain link diagram @xmath37 , that otherwise would have taken a quite amount of time to evaluate . the key is to work locally , that is , to cut the link into subtangles , compute the invariant for each subtangle and finally assembly the obtained invariants into the invariant of @xmath38 via the tensor product operation induced on formal complexes by the horizontal composition operation on the canopoly @xmath39 to really obtain an improvement of computational efficiency , we simplify the complexes over the category @xmath40 before assembling , using the _ delooping _ and _ gaussian elimination _ tools borrowed from @xcite but adapted to our geometric picture and local relations . lemma [ lemma : delooping ] below is similar to ( * ? ? ? * lemma 4.1 ) but uses our local relations , while lemma [ lemma : gaussian elimination ] is exactly ( * ? ? ? * lemma 4.2 ) therefore we omit its proof . [ lemma : delooping ] ( delooping ) given an object of the form @xmath41 in the category @xmath40 , where @xmath42{unknot - clockwise.pdf}}$ ] or @xmath43{circle2sv.pdf}}$ ] , it is isomorphic in mat(@xmath40 ) to the direct sum @xmath44 in which @xmath14 is removed . this can be written symbolically as @xmath45{unknot - clockwise.pdf}}\ , \cong\ , \emptyset \{+1\ } \oplus \emptyset \{-1\ } \quad \text{and } \quad \raisebox{-3pt}{\includegraphics[height=0.18in]{circle2sv.pdf } } \,\cong \ , \emptyset \{+1\ } \oplus \emptyset \{-1\}.\ ] ] the desired isomorphisms are given in figure [ delooping ] . let us first show that @xmath46{capor.pdf}},\raisebox{-5pt}{\includegraphics[height=.25in]{capordot.pdf}}\,)^t$ ] and @xmath47{cupordot.pdf } } -h \,\raisebox{-5pt}{\includegraphics[height=.25in]{cupor.pdf } } , \raisebox{-5pt}{\includegraphics[height=.25in]{cupor.pdf}}\,)$ ] are mutually inverse isomorphisms . @xmath48{cupordot.pdf } } -h \,\raisebox{-5pt}{\includegraphics[height=.25in]{cupor.pdf } } , & \raisebox{-5pt}{\includegraphics[height=.25in]{cupor.pdf}}\end{array } \right ) \circ \left ( \begin{array}{c}\raisebox{-5pt}{\includegraphics[height=.25in]{capor.pdf } } \vspace{.1 cm } \\\raisebox{-5pt}{\includegraphics[height=.25in]{capordot.pdf } } \end{array } \right ) = \raisebox{-15pt}{\includegraphics[height=0.5in]{surgery1.pdf } } - h\raisebox{-15pt}{\includegraphics[height=0.5in]{surgery3 } } + \raisebox{-15pt}{\includegraphics[height=0.5in]{surgery2.pdf } } \stackrel{(sf)}{= } \raisebox{-15pt}{\includegraphics[height=0.5in]{surgery.pdf } } = \operatorname{id}(\raisebox{-4pt}{\includegraphics[height=0.2in]{unknot - clockwise.pdf}})\ ] ] @xmath49{capor.pdf } } \vspace{.1 cm } \\\raisebox{-5pt}{\includegraphics[height=.25in]{capordot.pdf } } \end{array } \right ) \circ \left ( \begin{array } { cc } \raisebox{-5pt}{\includegraphics[height=.25in]{cupordot.pdf } } -h \,\raisebox{-5pt}{\includegraphics[height=.25in]{cupor.pdf } } , & \raisebox{-5pt}{\includegraphics[height=.25in]{cupor.pdf } } \end{array } \right ) = \left ( \begin{array}{cc } \raisebox{-5pt}{\includegraphics[width=0.25in]{sphd.pdf } } - h \,\raisebox{-5pt}{\includegraphics[width=0.25in]{sph.pdf } } & \raisebox{-5pt}{\includegraphics[width=0.25in]{sph.pdf } } \vspace{.15 cm } \\ \raisebox{-5pt}{\includegraphics[width=0.25in]{sph2d.pdf } } - h\ , \raisebox{-5pt}{\includegraphics[width=0.25in]{sphd.pdf } } & \raisebox{-5pt}{\includegraphics[width=0.25in]{sphd.pdf } } \end{array } \right ) \stackrel { ( s ) , ( 2d)}{= } \left ( \begin{array}{cc } 1 & 0 \\ 0 & 1 \end{array } \right ) .\ ] ] in the same manner , one can use relations ( ufo ) and ( cn ) to verify that @xmath50{capsa.pdf } } , \raisebox{-5pt}{\includegraphics[height=.25in]{capsad.pdf}}\,)^t$ ] and @xmath51{cupsad.pdf } } + hi\,\raisebox{-5pt}{\includegraphics[height=.25in]{cupsa.pdf } } , -i\ , \raisebox{-5pt}{\includegraphics[height=.25in]{cupsa.pdf}}\,)$ ] are also mutually inverse isomorphisms . ( gaussian elimination for complexes ) [ lemma : gaussian elimination ] if @xmath52 is an isomorphism in some additive category @xmath53 , then the complex segment in mat(@xmath53 ) @xmath54 \stackrel{\left ( \begin{array}{c } \alpha \\ \beta \end{array}\right)}{\longrightarrow } \left [ \begin{array}{c } b_1 \\ d \end{array } \right ] \stackrel { \left ( \begin{array}{cc } \phi & \delta \\ \gamma & \epsilon \end{array } \right)}{\longrightarrow } \left [ \begin{array}{c } b_2 \\ e \end{array } \right ] \stackrel{\left ( \begin{array}{cc } \mu & \nu \end{array}\right ) } { \longrightarrow } \left [ f \right ] ... \ ] ] is isomorphic to the complex segment @xmath54 \stackrel{\left ( \begin{array}{c } 0 \\ \beta \end{array}\right)}{\longrightarrow } \left [ \begin{array}{c } b_1 \\ d \end{array } \right ] \stackrel { \left ( \begin{array}{cc } \phi & 0 \\ 0 & \epsilon - \gamma \phi ^{-1 } \delta \end{array } \right)}{\longrightarrow } \left [ \begin{array}{c } b_2 \\ e \end{array } \right ] \stackrel{\left ( \begin{array}{cc } 0 & \nu \end{array}\right ) } { \longrightarrow } \left [ f \right ] ... \ ] ] the later is the direct sum of the contractible complex @xmath55 and the complex segment @xmath56 \stackrel{\left ( \beta \right)}{\longrightarrow } \left [ d \right ] \stackrel{\left ( \epsilon - \gamma \phi ^{-1 } \delta \right)}{\longrightarrow } \left [ e \right ] \stackrel{\left(\nu \right ) } { \longrightarrow } \left [ f \right ] ... \,$ ] . therefore , the first and last complex segments are homotopy equivalent . whenever an object in some formal complex @xmath57 contains an oriented loop or a basic closed web with two vertices , we remove it using lemma [ lemma : delooping ] , where @xmath58 is the homotopy subcategory of _ then we use lemma [ lemma : gaussian elimination ] to cancel all isomorphisms in the resulting complex . let us first see how one can use the tools described in section [ sec : fast computations ] to show the homotopy invariance of the complex @xmath10 $ ] under reidemeister moves . for this , one has to compute and simplify the complexes corresponding to each side of a certain reidemeister move , to obtain the same result for both sides . consider diagrams @xmath59{lkink.pdf}}\,\,\text{and}\ , d'=\raisebox{-13pt}{\includegraphics[height=0.4in]{reid1 - 1.pdf}}.$ ] the complex @xmath60 : \,\ , 0 \longrightarrow \underline{\left [ \raisebox{-8pt}{\includegraphics[height=0.3in]{reid1 - 2.pdf}}\right ] \{-1\}}\stackrel{\raisebox{-8pt}{\includegraphics[height=0.3in]{reid1-d.pdf}}}{\longrightarrow}\left [ \raisebox{-8pt } { \includegraphics[height=0.3in]{reid1 - 3.pdf}}\right ] \{-2\ } \longrightarrow 0 $ ] is isomorphic in @xmath58 to @xmath61{reid1 - 1.pdf } } \{-2\}\vspace{.1 cm } \\ \raisebox{-8pt}{\includegraphics[height=0.3in]{reid1 - 1.pdf}}\{0\ } \end{array } \right ] } \stackrel{\left ( \begin{array}{cc}\raisebox{-8pt}{\includegraphics[height=0.3in]{isom-8.pdf } } & \raisebox{-8pt}{\includegraphics[height=0.3in]{isom-8d.pdf } } \end{array } \right ) } { \longrightarrow}\left [ \raisebox{-8pt } { \includegraphics[height=0.3in]{reid1 - 3.pdf}}\right ] \{-2\ } \longrightarrow 0.\ ] ] note that the underlined objects above are at the cohomological degree @xmath62 the later complex decomposes into contractible complex ( its differential is an isomorphism ) @xmath63{reid1 - 1.pdf}}\right ] \{-2\}}\stackrel{\left ( \raisebox{-8pt}{\includegraphics[height=0.3in]{isom-8.pdf}}\right ) } { \longrightarrow } \left [ \raisebox{-8pt } { \includegraphics[height=0.3in]{reid1 - 3.pdf}}\right ] \{-2\ } \longrightarrow 0 \quad \text{and}\ ] ] @xmath64{reid1 - 1.pdf } } \,\right ] } \longrightarrow 0.\ ] ] hence , complexes @xmath65{lkink.pdf}}\,]$ ] and @xmath65{reid1 - 1.pdf}}\,]$ ] are homotopy equivalent . consider diagrams @xmath59{dreid2b.pdf}}$ ] and @xmath66{twoarcsop.pdf}}.$ ] the formal complex @xmath67 $ ] is the double complex given below , which is the tensor product of the formal complexes associated with the two crossings in @xmath68 @xmath69{d - sum.pdf } } \qquad \raisebox{-85pt}{\includegraphics[height=2.5in]{reid2-ch12.pdf}}\ ] ] @xmath70{chain - r2-ch12.pdf}\ ] ] the morphism is the ` singular saddle ' with domain and range , while the morphism is the ` singular saddle ' with domain and range . there is a loop in the previous complex , thus applying lemma [ lemma : delooping ] , @xmath67 $ ] is isomorphic in @xmath58 to : @xmath71{chain2-r2-ch12.pdf}}\ ] ] the above complex is the direct sum of @xmath72{chain - r2-component1.pdf}},\ ] ] @xmath73{chain - r2-component2.pdf } } \,\ , \ , \ , \text{and}\ ] ] @xmath73{chain - r2-component3.pdf}}.\ ] ] the last two complexes are contractible , since there differentials are isomorphisms in @xmath74 and the first complex is isomorphic to @xmath75 $ ] ( see lemma [ lemma : isomorphisms ] ) . removing contractible direct summands we obtain that @xmath67 $ ] and @xmath75 $ ] are homotopy equivalent . the other reidemeister moves have a similar approach . the figure eight knot diagram in figure [ figure eight knot ] is the connected sum of the two tangle diagrams @xmath76{t1.pdf}}$ ] and @xmath77{t2.pdf}}.$ ] @xmath78 : \left [ \raisebox{-8pt } { \includegraphics[height=.35in]{t1 - 2.pdf } } \right ] \{4\ } \stackrel{\left ( \begin{array}{c } \raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-2-up.pdf } } \vspace{.05 in } \\ \raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-2-down.pdf } } \end{array } \right ) } { \longrightarrow } \left [ \begin{array}{c } \raisebox{-8pt}{\includegraphics[height=.35in]{t1 - 1-up.pdf } } \vspace{.05 in } \\ \raisebox{-8pt}{\includegraphics[height=.35in]{t1 - 1-down.pdf } } \end{array } \right ] \{3\ } \stackrel{\left ( \begin{array}{cc } \raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-1-left.pdf } } & -\raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-1-right.pdf } } \end{array}\right)}{\longrightarrow } \underline{\left [ \raisebox{-8pt}{\includegraphics[height=.35in]{t10.pdf}}\right ] \{2\}}\ ] ] the object at height @xmath79 in @xmath80 $ ] contains a loop . applying lemma [ lemma : delooping ] and composing the morphisms in the second differential above with those of figure [ delooping ] we get the next complex , which is isomorphic to @xmath80 $ ] : @xmath81{t1 - 2.pdf } } \right ] \{4\ } \stackrel{\left ( \begin{array}{c } \raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-2-up.pdf } } \vspace{.05in}\\ \raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-2-down.pdf } } \end{array } \right ) } { \longrightarrow } \left [ \begin{array}{c } \raisebox{-8pt}{\includegraphics[height=.35in]{t1 - 1-up.pdf } } \vspace{.05in}\\ \raisebox{-8pt}{\includegraphics[height=.35in]{t1 - 1-down.pdf } } \end{array } \right ] \{3\ } \stackrel{\left ( \begin{array}{cc } \raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-1-fi.pdf } } & -\raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-1-delta.pdf } } \vspace{.05 in } \\ \raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-1-gama.pdf } } & -\raisebox{-5pt}{\includegraphics[height=.25in]{t1-d-1-epsilon.pdf}}\end{array}\right)}{\longrightarrow } \underline{\left [ \begin{array}{c}\raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf}}\{3\ } \vspace{.05in}\\ \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf}}\{1\ } \end{array } \right ] } \ ] ] for the simplicity of drawings , we next apply the isomorphisms [ isomorphisms ] to remove pairs of vertices . after this operation , the previous complex is isomorphic to : @xmath82{up - down.pdf } } \right ] \{4\ } \stackrel{\left ( \begin{array}{c } -i \raisebox{-5pt}{\includegraphics[height=.25in]{saddle - ud - lr.pdf } } \vspace{.05in}\\ -i \raisebox{-5pt}{\includegraphics[height=.25in]{saddle - ud - lr.pdf } } \end{array } \right ) } { \longrightarrow } \left [ \begin{array}{c } \raisebox{-8pt}{\includegraphics[height=.3in]{left - right.pdf } } \vspace{.05in}\\ \raisebox{-8pt}{\includegraphics[height=.3in]{left - right.pdf } } \end{array } \right ] \{3\ } \stackrel{\left ( \begin{array}{cc } \,\,\,\,\raisebox{-5pt}{\includegraphics[height=.25in]{left - right.pdf } } & -\raisebox{-5pt}{\includegraphics[height=.25in]{left - right.pdf } } \vspace{.05in}\\ -\raisebox{-5pt}{\includegraphics[height=.25in]{left - right - dup.pdf } } + h \,\raisebox{-5pt}{\includegraphics[height=.25in]{left - right.pdf } } & \raisebox{-5pt}{\includegraphics[height=.25in]{left - right - ddown.pdf } } - h\raisebox{-5pt}{\includegraphics[height=.25in]{left - right.pdf}}\end{array}\right)}{\longrightarrow } \underline{\left [ \begin{array}{c}\raisebox{-8pt}{\includegraphics[height=.3in]{left - right.pdf}}\{3\ } \vspace{.05in}\\ \raisebox{-8pt}{\includegraphics[height=.3in]{left - right.pdf}}\{1\ } \end{array } \right ] } \ ] ] when appearing as a cobordism , the symbol denotes the identity automorphism of the resolution with the same symbol , that is , it is the union of two ` curtains ' . similarly , and denote the same cobordism with an extra dot on one of the curtains . moreover , denotes the saddle with domain and range . the upper left entry in the second nontrivial differential of the previous complex is an isomorphism , and applying the first part of lemma [ lemma : gaussian elimination ] we arrive at the complex below : @xmath81{up - down.pdf } } \right ] \{4\ } \stackrel{\left ( \begin{array}{c } 0 \vspace{.05in}\\ -i \raisebox{-5pt}{\includegraphics[height=.25in]{saddle - ud - lr.pdf } } \end{array } \right ) } { \longrightarrow } \left [ \begin{array}{c } \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf } } \vspace{.05in}\\ \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf } } \end{array } \right ] \{3\ } \stackrel{\left ( \begin{array}{cc } \raisebox{-5pt}{\includegraphics[height=.25in]{left - right.pdf } } & 0 \vspace{.05in}\\ 0 & \raisebox{-5pt}{\includegraphics[height=.25in]{left - right - ddown.pdf } } - \raisebox{-5pt}{\includegraphics[height=.25in]{left - right - dup.pdf}}\end{array}\right)}{\longrightarrow } \underline{\left [ \begin{array}{c}\raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf}}\{3\ } \vspace{.05in}\\ \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf}}\{1\ } \end{array } \right ] } \ ] ] removing the contractible summand @xmath83{left - right.pdf}}\right ] \{3\ } \stackrel { \left ( \raisebox{-5pt}{\includegraphics[height=.25in]{left - right.pdf } } \right ) } { \longrightarrow } \underline{\left [ \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf } } \right ] \{3\}}\longrightarrow 0,\ ] ] we obtain the complex @xmath84 given below , which is isomorphic to @xmath80 $ ] in @xmath85 @xmath86{up - down.pdf } } \right ] \{4\ } \stackrel { \left ( -i \,\raisebox{-5pt}{\includegraphics[height=.25in]{saddle - ud - lr.pdf } } \right ) } { \longrightarrow } \left [ \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf}}\right ] \{3\ } \stackrel { \left ( \,\raisebox{-5pt}{\includegraphics[height=.25in]{left - right - ddown.pdf } } - \,\raisebox{-5pt}{\includegraphics[height=.25in]{left - right - dup.pdf}}\right ) } { \longrightarrow } \underline{\left [ \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf}}\right ] \{1\}}\ ] ] the complex @xmath87 $ ] associated to the other half of the figure eight knot diagram is computed and simplified similarly . it turns out that it is isomorphic in @xmath58 to the complex @xmath88 depicted below . @xmath89{up - down.pdf } } \right ] \{-1\ } } \stackrel { \left ( i \,\raisebox{-5pt}{\includegraphics[height=.25in]{up - down - dleft.pdf } } -i \,\raisebox{-5pt}{\includegraphics[height=.25in]{up - down - dright.pdf } } \right ) } { \longrightarrow } \left [ \raisebox{-8pt}{\includegraphics[height=.35in]{up - down.pdf}}\right ] \{-3\ } \stackrel { \left ( -\,\raisebox{-5pt}{\includegraphics[height=.25in]{saddle - ud - lr.pdf } } \right ) } { \longrightarrow } \left [ \raisebox{-8pt}{\includegraphics[height=.35in]{left - right.pdf}}\right ] \{-4\}\ ] ] next step is to take the ` formal tensor product ' of @xmath84 with @xmath88 using the same side - by - side composition that one has to use to get from @xmath90 and @xmath91 the figure eight knot diagram . as a result , we obtain the double complex @xmath53 shown below , in which we smoothed out the resolutions and cobordisms ; we also canceled the four morphisms obtained on the upper right of the diagram , as they are differences of the same cobordism . @xmath92{complex - c.pdf}\ ] ] the following step is to replace every loop with a pair of degree - shifted empty sets as in lemma [ lemma : delooping ] , and to replace the differentials with their compositions with the isomorphisms of figure [ delooping ] . as every object of @xmath53 contains only loops , we arrive at the complex @xmath93 depicted below , in which all objects are degree - shifted empty sets and all morphisms are matrices of scalar multiples of the empty cobordism recall that we are working modulo the local relations @xmath27 and all closed foams reduce to an element of the ground ring @xmath2.$ ] let us recall from @xcite the structure maps for the frobenius algebra defined on @xmath94/(x^2 -h x -a ) = { \langle 1 , x\rangle}_{\mathbb{z}[i , a , h]}.$ ] the unit map @xmath95[a , h ] \rightarrow \mathcal{a}$ ] and counit map @xmath96[a , h]$ ] are given by @xmath97 and @xmath98 respectively . multiplication @xmath99 and comultiplication @xmath100 are defined by @xmath101 we will use the basis @xmath102 of the algebra @xmath103 and write the maps @xmath104 and @xmath105 relative to this basis . the cobordism is the multiplication by @xmath106 endomorphism of @xmath103 on the first component of @xmath107 and the identity map on the second component . likewise , is the identity on the first component and multiplication by @xmath106 endomorphism of @xmath103 on the second component of the tensor product . therefore , these cobordisms are defined by the following rules : @xmath108{two - circles - leftd.pdf } } & = \left \ { \begin { array}{ccccc } 1 \otimes 1 & \rightarrow & x \otimes 1 \\ 1 \otimes x & \rightarrow & x \otimes x \\ x \otimes 1 & \rightarrow & x ^2\otimes 1 & = & h ( x\otimes 1 ) + a ( 1\otimes 1)\\ x \otimes x & \rightarrow & x ^2\otimes x & = & h(x \otimes x ) + a ( 1 \otimes x ) \end{array } \right . \\ \raisebox{-5pt}{\includegraphics[height=.25in]{two - circles - rightd } } & = \left \ { \begin { array}{ccccc } 1 \otimes 1 & \rightarrow & 1 \otimes x \\ 1 \otimes x & \rightarrow & 1 \otimes x^2 & = & h ( 1 \otimes x)+ a ( 1\otimes 1 ) \\ x \otimes 1 & \rightarrow & x \otimes x \\ x \otimes x & \rightarrow & x \otimes x^2 & = & h ( x\otimes x ) + a ( x \otimes 1 ) \end{array } \right.\end{aligned}\ ] ] the matrix of the cobordism @xmath109{two - circles - leftd.pdf } } -i \raisebox{-5pt}{\includegraphics[height=.25in]{two - circles - rightd.pdf}}$ ] relative to the basis @xmath110 of the tensor product @xmath107 is : @xmath111{two - circles - leftd.pdf } } -i \,\raisebox{-5pt}{\includegraphics[height=.25in]{two - circles - rightd.pdf } } = \ \ \left ( \begin{array}{cccc } 0 & -ai & ai & 0 \\ -i & -hi & 0 & ai \\ i & 0 & hi & -ai \\ 0 & i & -i & 0 \end{array } \right ) .\ ] ] likewise we have @xmath112{twoconc - circles - upd.pdf } } - \raisebox{-5pt}{\includegraphics[height=.25in]{twoconc - circles - downd.pdf } } = \ \ \left ( \begin{array}{cccc } 0 & -a & a & 0 \\ -1 & -h & 0 & a \\ 1 & 0 & h & -a \\ 0 & 1 & -1 & 0 \end{array } \right ) .$ ] we are now ready to write the complex @xmath113 @xmath114{complex - lambda1.pdf}}\ ] ] there are many isomorphisms in @xmath115 thus we repeatedly apply lemma [ lemma : gaussian elimination ] until no invertible entries remain in any of the matrices . adding relations @xmath116 and working over @xmath117 ( note that we can also use here @xmath118 any non - zero number is invertible . henceforth we obtain the double complex @xmath119 given below , in which all matrices are @xmath62 @xmath120{complex - lambda2.pdf}}\ ] ] taking the total complex of @xmath119 we arrive at the complex @xmath121 which is homotopy equivalent to @xmath122,$ ] where @xmath123 is the figure eight knot diagram we started with : @xmath124{complex - lambda3.pdf}}.\ ] ] we need now to apply the functor @xmath125 defined in @xcite , to obtain an ordinary complex with objects graded vector spaces over @xmath117 and take its cohomology . since @xmath126 we have : @xmath127 computing the cohomology of the complex @xmath128 we obtain that the cohomology group over @xmath117 of the figure eight knot is 6dimensional , with generators in bidegrees @xmath129 , @xmath130 and @xmath131 that is , after adding the relations @xmath116 , the cohomology groups of the figure eight knot are : @xmath132 } \mathbb{c}= \begin{cases}\mathbb{c } \ \ \ \ \text{for}\ \ \ ( i , j ) \in \{(-2,5 ) , ( -1,1 ) , ( 0 , -1 ) , ( 0,1 ) , ( 1 , -1),(2,-5)\ } \\ 0 \ \ \ \ \text{otherwise}. \end{cases}\ ] ] when implementing the delooping and gaussian elimination tools , there is a better way than computing the ` half knot ' invariants and putting the results together , as noted in ( * ? ? ? * section 7 ) . instead , one picks a certain crossing of the knot and adds one crossing at a time . after each crossing is added , the two tools are used to simplify the result before moving to the next crossing . in this section we describe a purely topological variant of our universal @xmath0 dotted foam theory , in the sense that no dots are present on singular cobordisms . however , one has to pay a price for it , because this theory works if @xmath133 exists in the ground ring . moreover , we will also lose the beautiful geometric interpretation of certain algebra structure properties on @xmath103 . for example , relation ( 2d ) is the geometric counterpart of the identity @xmath134 in @xmath135 and the surgery formula ( sf ) corresponds to @xmath136 both of them are lost . the main idea is to use the _ genus reduction _ formula from @xcite , which for our purpose here , we write it in the following form : @xmath137{capordot.pdf } } = \frac{1}{2 } ( \ , \raisebox{-8pt}{\includegraphics[height=.3in]{cap - g1.pdf } } + h\ , \raisebox{-8pt}{\includegraphics[height=.3in]{capor.pdf}}\ , ) .\ ] ] we consider the ground ring @xmath138 $ ] and the @xmath139-module @xmath140/(x^2 - hx -a)$ ] with generators @xmath141 and @xmath142 on which we consider the same frobenius structure maps we used for @xmath143 the local relations @xmath27 = ( 2d , sf , s , ufo ) from @xcite are replaced by @xmath144 given below . note that the _ genus two reduction _ formula @xmath145 is the replacement of the _ two dots _ relation ( 2d ) , and that the _ torus identity _ @xmath146 corresponds to the second identity in ( s ) . @xmath147{surgery.pdf } } = \displaystyle\frac{1}{2 } \raisebox{-22pt}{\includegraphics[height=.7in]{surgery1-g1.pdf } } + \frac{1}{2}\raisebox{-22pt}{\includegraphics[height=.7in]{surgery2-g1.pdf } } \hspace{2 cm } \raisebox{-10pt}{\includegraphics[width=0.4in]{sph.pdf}}=0 & ( \tilde{s } ) \\ \raisebox{-13pt}{\includegraphics[height=.4in]{ufo - g0.pdf } } = 0 = \raisebox{-13pt}{\includegraphics[height=.4in]{ufo - g2.pdf}}\ , , \quad \raisebox{-13pt}{\includegraphics[height=.4in]{ufo - g1-lower.pdf } } = 2i\ , , \quad \raisebox{-13pt}{\includegraphics[height=.4in]{ufo - g1-upper.pdf } } = -2i & ( \tilde{ufo})\\ \ ( g2)\ \ \ \ \ \ \raisebox{-8pt}{\includegraphics[height=.35in]{cap - g2.pdf } } = ( h^2 + 4a)\ , \raisebox{-8pt}{\includegraphics[height=.35in]{capor.pdf } } \hspace{2.3 cm } \raisebox{-5pt}{\includegraphics[width=0.6in]{torus.pdf}}=2 & ( t ) } \ ] ] the _ curtain identities _ still hold , as well as the isomorphisms given in lemma [ lemma : isomorphisms ] . with the new relations @xmath148 the _ cut - neck relation _ ( cn ) and identity ( rsc ) become @xmath149 and @xmath150 respectively , which are depicted below . @xmath151{cneck0.pdf}}= \displaystyle -\frac{i}{2 } \raisebox{-25pt}{\includegraphics[width=0.35 in , height=0.85in]{cneck1-g1.pdf } } - \frac{i}{2 } \raisebox{-25pt}{\includegraphics[width=0.35 in , height=0.85in]{cneck2-g1.pdf } } & ( \tilde{cn } ) \\ \raisebox{-22pt}{\includegraphics[height=0.7in]{remsc.pdf } } = \displaystyle\frac{i}{2 } \raisebox{-22pt}{\includegraphics[height=0.7in]{remsc1-g1.pdf } } - \frac{i}{2 } \raisebox{-22pt}{\includegraphics[height=0.7in]{remsc2-g1.pdf } } & ( \tilde{rsc } ) } \ ] ] interestingly , relations [ exchanging dots ] are replaced here by : @xmath152{cupsa - g1.pdf } } + \raisebox{-8pt}{\includegraphics[height=.35in]{cupsa - g1-right.pdf } } = 0 \hspace{2 cm } \raisebox{-8pt}{\includegraphics[height=.35in]{cupsa - g2.pdf } } = -(h^2 + 4a)\ , \raisebox{-8pt}{\includegraphics[height=.35in]{cupsa.pdf}}\end{aligned}\ ] ] therefore , handles can be exchanged between two neighboring facets of a foams , at the expense of a minus sign . we denote by @xmath153 the category of dotless foams modulo the new local relations @xmath154 the tautological functor @xmath155-mod is defined on links as before , namely on objects is defined by @xmath156 and on morphisms by composition on the left . using the isomorphisms from lemma [ lemma : isomorphisms ] , any resolution of a link diagram is isomorphic in @xmath153 with a disjoint union of oriented loops and closed webs with two vertices . note that also a closed web with two vertices is isomorphic in @xmath153 to an oriented loop . repeatedly applying the local relations @xmath157 we get to foams in which every component has one boundary curve ( which is either or ) and is either of genus @xmath79 or of genus @xmath158 in particular , @xmath159{unknot - clockwise.pdf}})$ ] is an @xmath139-module generated by @xmath160{cupor.pdf}}$ ] and @xmath161{cupor - g1.pdf } } + \frac{1}{2 } h\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}$ ] and @xmath162{circle2sv.pdf}})$ ] is an @xmath139-module generated by @xmath163{cupsa.pdf}}$ ] and @xmath164{cupsa - g1.pdf } } + \frac{1}{2 } h \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupsa.pdf}}.$ ] it is easy to see that there are degree - preserving @xmath139-module isomorphisms @xmath159{unknot - clockwise.pdf } } ) \cong \mathcal{a } \cong \mathcal{f}(\raisebox{-3pt}{\includegraphics[height=.16in]{circle2sv.pdf}})$ ] that map @xmath165 to @xmath141 and @xmath166 to @xmath167 moreover , one can verify that with this basis for @xmath159{unknot - clockwise.pdf}})$ ] the tautological functor @xmath168 behaves on links in the same manner as the tqft functor @xmath169 $ ] defined by the frobenius system corresponding to @xmath170 where @xmath171 is the tqft functor used in @xcite . furthermore , @xmath169 $ ] satisfies relations @xmath154 for example , to show that @xmath169 $ ] satisfies relation @xmath145 we need to check @xmath172 ( recall that we read cobordisms from bottom to top ) . for @xmath173 we need to verify that @xmath174 both equalities hold . moreover , @xmath169 $ ] satisfies relations @xmath175 and @xmath146 since @xmath176 and @xmath177 finally , a singular circle corresponds to multiplication by @xmath178 endomorphism of @xmath179 ( see @xcite ) , and we let the details of checking that @xmath169 $ ] satisfies relations @xmath180 to the reader . on the other hand we have : @xmath137{pair - of - pants.pdf } } \circ ( \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \otimes \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ , ) = \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ ] ] @xmath137{pair - of - pants.pdf } } \circ [ \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \otimes ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2}\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \ , ) ] = \frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } = \raisebox{-8pt}{\includegraphics[height=.3in]{pair - of - pants.pdf } } \circ [ ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2}\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ , ) \otimes \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \,]\ ] ] @xmath137{pair - of - pants.pdf } } \circ [ ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2}\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ , ) \otimes ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2}\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ , ) ] = h\,(\ , \frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2}\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ , ) + a\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ ] ] moreover , @xmath182{capor.pdf } } \circ \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } = \raisebox{-5pt}{\includegraphics[height=.2in]{sph.pdf } } \ \ \ \text{and } \ \ \ \ \raisebox{-5pt}{\includegraphics[height=.2in]{capor.pdf } } \circ ( \,\frac{1}{2 } \ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{1}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \ , ) = \frac{1}{2}\ , \raisebox{-3pt}{\includegraphics[height=.15in]{torus.pdf } } + \frac{h}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{sph.pdf}}\ ] ] @xmath137{upsidedown - pants.pdf } } \circ \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \stackrel{(\tilde{sf})}{= } \frac{1}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } \ \ \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } + \frac{1}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \ \ \raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } = ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2}\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \,)\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } + \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ , ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2}\ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \ , ) - h \ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \ , \raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ ] ] @xmath137{upsidedown - pants.pdf } } \circ ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \ , ) \stackrel{(\tilde{sf } ) , ( g2)}{= } ( \ , \frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ , ) \ , ( \,\frac{1}{2}\,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor - g1.pdf } } + \frac{h}{2 } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \ , ) + a \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf } } \,\raisebox{-5pt}{\includegraphics[height=.2in]{cupor.pdf}}\ ] ]

we show how to use bar - natan s ` divide and conquer ' approach to computation to efficiently compute the universal @xmath0 dotted foam cohomology groups , even for big knots and links . we also describe a purely topological version of the @xmath0 foam theory , in the sense that no dots are needed on foams .

fega@xmath0 is a rare intermetallic insulator that has attracted particular attention due to its unusual transport and magnetic properties . this fe - based material exhibits semiconducting behavior with a gap of 0.5 ev obtained from transport measurements.@xcite particular attention has been paid to understand the mechanism of the gap formation in the context of strong hybridization between fe-@xmath2 and ga-@xmath4 orbitals , reminiscent of that in strongly correlated @xmath2 and @xmath5 kondo insulators.@xcite given the small band gap and the presence of very narrow bands around the fermi level , fega@xmath0 has been most extensively studied as a thermoelectric material . high values of the seebeck coefficient around 350 @xmath6v / k have been measured at room temperature.@xcite in single crystals colossal values of the thermopower ( @xmath7 -16000 @xmath6v / k ) emerge below 20 k due to the phonon drag effect.@xcite there is no unambiguous picture about the role of electronic correlations and the magnetic ( or not ) character of stoichiometric fega@xmath0 from experiments . according to susceptibility measurements , fega@xmath0 is diamagnetic below room temperature ( rt ) and the susceptibility shows an increase above rt suggesting proximity to a crossover to a paramagnetic metallic state.@xcite fe mssbauer spectra did not show the presence of an internal magnetic field at the fe site , supporting a nonmagnetic state of fe @xcite though not ruling out correlated states . in contrast , muon spin rotation studies detected spectroscopic features interpreted in terms of electron confinement into spin polarons that require the existence of fe moments.@xcite the narrow gap in fega@xmath0 and its unusual properties suggesting electronic correlations brings to mind related iron compounds fesi@xcite and fesb@xmath1,@xcite whose underlying electronic systems remain to be fully understood . hole or electron doping drastically changes the properties of the parent compound , giving rise to emergent magnetic phases . fega@xmath0 has a very distinctive structural characteristic : fe atoms occur in dimers as shown in fig . [ struct ] . replacing fe breaks the dimer symmetry , a fundamental impact _ if _ there is important fe - fe bonding , spin correlations , or singlet formation . focusing first on electron doping , co - substitution for fe induces an insulator - to - metal transition . @xcite resistivity @xmath8 measurements for fe@xmath9co@xmath10ga@xmath0 are not yet conclusive ; some indicate the metallic state is reached at doping level @xmath11=0.025 - 0.075,@xcite in other reports only at substantially higher doping levels of @xmath11=0.125@xcite or 0.23.@xcite analysis of the t - dependence of the nuclear spin - lattice relaxation rate 1/t@xmath12 of the @xmath13ga nuclei suggests the existence of in - gap states at low co doping.@xcite in fe@xmath14co@xmath14ga@xmath0 , the relaxation is strongly enhanced due to spin fluctuations , often a signature of a weakly antiferromagnetic ( afm ) metal . such itinerant antiferromagnetic behavior contrasts with magnetization measurements , showing localized magnetism with a relatively low effective moment of 0.7@xmath15/f.u.@xcite electron - doping by substituting ga by ge leads to drastically different behavior . @xcite experiments show that fega@xmath16ge@xmath17 is conducting at an extremely low doping level @xmath18= 0.0006 , progressing to a weak ferromagnetic order at @xmath18@xmath19= 0.13 which never appears in the co - doped compound . the emergence of the ferromagnetic ( fm ) state is accompanied by quantum critical behavior observed in the specific heat and the magnetic susceptibility . the fm instability found in fega@xmath16ge@xmath17 beyond @xmath18@xmath20= 0.13 indicates that strong electron correlations are induced by the disturbance of fe @xmath2 ga @xmath4 hybridization , or possibly that existing strong correlations are disrupted . @xcite turning now to hole doping , gamza _ _ et al.__@xcite performed resistivity and thermodynamic measurements on single crystals of fega@xmath0 , fe@xmath9mn@xmath10ga@xmath0 and fega@xmath16zn@xmath17 ( x @xmath21 0.12 and y @xmath21 0.06 ) . unlike for electron doping , hole doping using mn on the fe site or zn on the ga site does not give rise to a semiconductor - to - metal transition . hole doping induces states into the semiconducting gap that remain localized at the highest doping levels . using neutron powder diffraction measurements , they conclude that fega@xmath0 orders magnetically above room temperature in a complex structure , unaffected by the inclusion of mn and zn . evidently input from theoretical modeling is required to move toward understanding of this unusual behavior . one emphasis of this paper is to give special attention to the metal dimers in this structure . the flat bands bordering the gap suggest correlated electron behavior , and interatomic correlations should be much stronger between electrons or spins within a dimer than between dimers , and local physics may appear . the situation can be illustrated by considering fe@xmath14co@xmath14ga@xmath0 mentioned above . any given dimer has a 25% chance of being fe@xmath1 , 25% chance of being co@xmath1 , and 50% chance of feco . these three configurations may have very different physics , and their states may lie in the gap ( and may destroy it ) or in the sea of itinerant states . the system average may be very complex . doping on the ga site , however , does not directly disturb the dimers , and the change in band filling may be easier to handle and to understand . dft - based calculations for fega@xmath0 have been reported using the two more common ( semi)local approaches for the exchange correlation energy and potential : local density approximation ( lda)@xcite and generalized gradient approximation ( gga).@xcite the very similar results seem to be sufficient for describing the electronic structure of the undoped compound , giving a value@xcite of the gap of 0.4 - 0.5 ev comparable to the experiments . however , those approaches are unable to model the afm phase in co - doped fega@xmath0 , and they incorrectly predict a metallic fm ground state for hole doped fega@xmath0 , whether by ga or fe substitution . within the correlated dft lda+@xmath22 method ( see sec . vi ) , local moments arise on the fe dimer if the spins are antialigned , giving an antiferromagnetic result in band theory . in view of these results , yin and pickett@xcite suggested that the fe dimers could be forming spin singlets strongly correlated but nonmagnetic local states and that magnetism found in doped fega@xmath0 would be linked to the breaking of the singlets into free spins . singh showed @xcite that the magnetism of some types of doping of fega@xmath0 can be explained within gga without the need of spin coupling of pre - existing moments . this weakly correlated picture suggests that both @xmath23-type and @xmath24-type fega@xmath0 will become itinerant ferromagnets@xcite due to the large density of states on either side of the gap . recently , it has been proposed from dft - based calculations combined with fe moessbauer spectroscopy that increasing ge - doping level ( @xmath18 ) in fega@xmath16ge@xmath17 over a wide range 0.03@xmath21@[email protected] leads to an evolution from localized moments to a combination of localized and itinerant moments with the interplay of ferromagnetism and antiferromagnetism.@xcite dynamical mean field theory ( dmft ) calculations have also been performed in fega@xmath0,@xcite showing that fe ions are dominantly in an @xmath251 state displaying strong spin and charge fluctuations . the dmft approach is unable to model interatomic spin correlations . the use of a non - multiplicative ( non - local ) potential such as lda+@xmath22 and the explanation in terms of an antiferromagnetic ground state for undoped fega@xmath0@xcite introduces new questions about the role of strong electronic correlations . in this paper we revisit several aspects of the electronic structure of electron and hole - doped fega@xmath0 , comparing results from the lda and lda+@xmath22 functional forms . because in several cases the value of the gap is of some importance , in the appendix we indicate how the modified becke - johnson potential ( see within ) shifts bands and in some cases modifies magnetic moments giving better agreement with the experimental results . fega@xmath0 crystallizes in the tetragonal space group p4@xmath1/mnm . its lattice constants are @xmath26= 6.2628 and @xmath27= 6.5546 with four formula units ( two fe@xmath1 dimers ) per unit cell . fe atoms are at ( 0.3437 , 0.3437 , 0 ) and form dimer pairs in the @xmath28=0 plane along ( 110 ) and in @xmath28= 1/2 along the @xmath29 directions , as illustrated in fig . [ struct ] . there are two inequivalent ga sites : higher symmetry ga1 at ( 0 , 0.5 , 0 ) , and lower symmetry ga2 at ( 0.1556 , 0.1556 , 0.262 ) . each fe atom has eight ga neighbors , two ga1 at distances of 2.36 , and six ga2 , two of them at 2.39 and four at 2.48 ( see fig . [ struct ] ) . the paired fe atoms are separated by 2.77 , 12 @xmath30 larger distance than that between fe atoms in bcc fe metal ( 2.48 ) . ( color online ) upper panel : crystal structure of the unit cell of fega@xmath0 . fe atoms form dimers oriented along the [ 110 ] and [ 1@xmath310 ] directions . the local coordinate system used in the dos plots with the local @xmath28 axis directed along the fe@xmath1 dimers is circled . lower panel : local environment of fe atoms in a dimer formed by ga2 at 2.39(2 ) and 2.48(4 ) and ga1 at 2.36(2 ) . the two distinct sites for ga atoms can be clearly identified . the local coordinate system used in the density of states plots is also shown . the unit cell shown corresponds to a structure with space group _ pm _ ( lower symmetry than the experimental one , see section iii).,title="fig : " ] ( color online ) upper panel : crystal structure of the unit cell of fega@xmath0 . fe atoms form dimers oriented along the [ 110 ] and [ 1@xmath310 ] directions . the local coordinate system used in the dos plots with the local @xmath28 axis directed along the fe@xmath1 dimers is circled . lower panel : local environment of fe atoms in a dimer formed by ga2 at 2.39(2 ) and 2.48(4 ) and ga1 at 2.36(2 ) . the two distinct sites for ga atoms can be clearly identified . the local coordinate system used in the density of states plots is also shown . the unit cell shown corresponds to a structure with space group _ pm _ ( lower symmetry than the experimental one , see section iii).,title="fig : " ] the electronic structure calculations were performed with the wien2k code,@xcite based on density functional theory ( dft ) utilizing the augmented plane wave plus local orbitals method ( apw+lo).@xcite the calculations were well converged with respect to @xmath32mesh and r@xmath33k@xmath34 , using up to 1000 k - points ( 10@xmath3510@xmath359 mesh ) and r@xmath33k@xmath34=7.0 cutoff . selected sphere radii ( a.u . ) were the following : 2.27 for fe , 2.16 for ga , 2.35 for mn and co , 2.16 for ge , and 2.22 for zn . because the description of the electronic structure is in question , an assortment of exchange - correlation potentials have been used : lda,@xcite lda+@xmath22 ( lda plus the on - site repulsion @xmath22 within the fully localized limit functional),@xcite and the tran - blaha modified becke johnson ( from hereon mbj for simplicity ) included with lda . the lda+@xmath22 scheme improves over gga or lda in the study of systems containing strong intraatomic repulsion such as occurs in many transition metal compounds . @xcite we have calculated the effective @xmath22 for fega@xmath0 using the approach proposed by madsen and novak@xcite for augmented plane wave methods based on the procedure of anisimov and gunnarsson.@xcite in this method the occupation of a target orbital ( in this case the 3@xmath36 orbitals of fe ) is enforced . due to the ambiguity in the charge state of fe ( see discussion below ) , different 3@xmath36 occupations from @xmath37 to @xmath38 were presumed , with resulting effective @xmath22 values increasing linearly from 2 to 5 ev , respectively . we have performed calculations using the fully localized limit double counting functional within this @xmath22 range , with @xmath39 being set to 0.8 ev . the same value of @xmath22 has been applied simultaneously to fe , and to neighboring atoms mn and co. the mbj exchange potential ( a local approximation to an atomic exact - exchange potential and a screening term ) + lda correlation allows the calculation of band gaps with an accuracy similar to the much more expensive gw or hybrid methods . @xcite we have also studied the electronic structure of both undoped and doped fega@xmath0 by using the mbj potential which does not contain any system - dependent parameter . calculations with the magnetic moments within the metal dimers being aligned and antialigned have been performed . to be able to establish different magnetic orderings , a lower symmetry structure ( with space group _ pm _ ) was used . this structure contains 12 inequivalent atoms : four inequivalent fe sites ( fe1-fe2 forming one dimer and fe3-fe4 forming the other dimer ) , four inequivalent ga1 sites , and four inequivalent ga2 sites ( see fig . [ struct ] ) . table [ table1 ] shows , for each of the doping mechanisms and computational schemes studied , the magnetic moments of each of the atoms in the metal dimers , the total magnetic moment in the cell , and the band gap for the corresponding magnetic ground state . [ cols="<,^,^,^,^,^ " , ] within lda , illustrating the predicted band gap that is in agreement with experiment . right panel : orbital resolved dos for fe orbitals , in local coordinates ( see text and fig . [ struct ] ) for nonmagnetic fega@xmath40 and obtained within lda . the fermi level is at zero . ] _ lda . _ both ferromagnetic and antiferromagnetic orientations within the fe@xmath1 dimers have been studied . within lda ( gga ) only a nonmagnetic state can be stabilized ( see table [ table1 ] ) . this result is consistent with the observed diamagnetism in pure fega@xmath0 and with previous electronic structure calculations.@xcite in fig . [ dos_lda ] the band structure with band character plot and dos of fega@xmath0 within lda are shown . the electronic state is insulating with a band gap of 0.52 ev . flat bands both below and above the gap give rise to a large dos into which electrons or holes will be doped . most of the contribution to states around the fermi level comes from narrow fe-@xmath36 bands with little ga-@xmath24 contribution . for fe @xmath2 states we use a local coordinate system with the local @xmath28 axis oriented along the dimer axis ( as shown in fig . [ struct ] ) . using this local coordinate system the dos looks as shown in the right panel of fig . [ dos_lda ] with seemingly pseudo - cubic @xmath41 , @xmath42 , @xmath43 occupation . below the gap , occupied @xmath44 bands form a narrow , 0.35 ev wide four - band / spin complex . above the gap , @xmath45 and @xmath46 bands form a 0.4 ev wide two - band / spin complex . this different orbital character had been pointed out earlier by singh@xcite and by yin and pickett.@xcite _ lda+u . _ for the undoped compound , when an onsite coulomb repulsion is included , a magnetic state can only be obtained for an afm ordering within the dimers at low @xmath22 ( the magnetic moments vanish when fm order is set ) . the magnitude of the magnetic moment increases gradually as the @xmath22 value does ( from 0.5@xmath15 for @xmath47 2 ev , to 1.07@xmath15 for @xmath47 3ev , and 1.63@xmath15 for @xmath47 4 ev ) . the derived magnetic moments are in agreement with those obtained by yin _ _ et al.__@xcite an insulating state is retained for @xmath22 in the 2 - 4 ev range , but surprisingly the band gap closes as @xmath22 increases . the antiferromagnetic spin singlet scenario derived from lda+@xmath22 calculations at @xmath22= 3 ev gives the same band gap as the lda result ( with the doss differing only in some specifics ) and is consistent with transport and thermodynamic experiments ( see table [ table1 ] ) . in halides , oxides , and some chalcogenides of iron , it is possible and very useful to identify the charge state ( also known as formal valence ) of fe . this underlying picture provides substantial guidelines on the character of excitations that are likely to dominate the low energy behavior of the system . in some unusual ( semi)metallic transition metal compound , viz . cosb@xmath0,@xcite such identification proves to be difficult , with unconventional pictures arising . we preface our study of doping of fega@xmath0 with information relating to the fe charge state ( also called formal valence ) . we remind that the formal charge state often has only a very indirect connection to the physical charge density of atoms in the compound.@xcite the integrated partial dos , obtained from projections of the bloch states , provides a guideline for the @xmath2 occupation of fe . supposing that true @xmath2 character falls off above 1.5 ev , ( i.e. higher lying @xmath36 character reflecting tails of ga atoms extending into the fe sphere ) , the @xmath41 , @xmath42 , @xmath43 orbitals are fully occupied , 6 e@xmath48/fe . the other two orbitals , local @xmath45 and @xmath46 ( which are very distinct orientationally and probably chemically ) , are 75% occupied , thus providing 3 @xmath49/fe . the inferred occupation is then a surprising fe@xmath50 : @xmath51 . this characterization would be ( fe@xmath1)@xmath52 ( ga@xmath53)@xmath54 . although this characterization is not outrageous ga is quite electropositive and easily donates electrons to neighboring electronegative atoms , this picture should be given further scrutiny . there are , incontrovertibly , 17 occupied bands per spin channel per fe@xmath1 dimer cluster . this situation , and especially the gap obtained even in the weakly correlated treatment , strongly suggests covalent ( or metallic ) bonding rather than ionic bonding as in many fe insulators . since the starting procedure is to assign orbitals localized at specific sites in the initial projection to obtain wannier functions ( wfs),@xcite our choice was to have ( per spin ) 4 electrons per fe , two electrons for six of the ga atoms and one electron for the other six ga atoms . often in complex structured materials these initial associations do not persist , and that was the case here . the agreement between the occupied band structure obtained from wannier function interpolation and that derived from the dft calculation is excellent ( see fig . [ bs_wannier ] ) , indicating a faithful ( though not unique ) transformation to wfs . four different initial projections resulted in four sets of wfs with the same total spread ( mean localization ) . a persistent feature throughout the sets was that , of the 34 wfs / spin , precisely two are centered in the middle of dimers , i.e. one per dimer as can be seen in the top left panel of fig . [ wf ] , or 1/2 electron per fe . most of the other 32 wannier functions / spin resemble each other : lopsided objects with much of the occupation assignable to the fe site ( top right panel of fig.[wf ] ) . in some cases 1 - 3 @xmath56-like wfs centered on fe result , and these have the lowest spread of any of the wfs ( lower panel of fig . [ wf ] ) . the `` dimer bonding wf '' introduces a peculiar and , to our knowledge , unprecedented local orbital picture of an unpolarized metal - atom - based insulator , and is related directly to the dimer structure in the cell . it corresponds to a one - center two - electron bond between the two fe atoms , and is directly analogous to the h@xmath1 molecule ( where the wf density is equivalent to the total density per spin since there is only one orbital per spin ) . since the character of the density just below the gap is antibonding , as is clear from fig . 5 , and the fe @xmath2 shell is more than half filled , this bonding dimer wf must arise primarily from deeper , bonding fe states . -mesh used in wannier function calculation does not include @xmath57-points along some symmetry lines . therefore , that the interpolated band structure from wannier functions agrees with the bloch bands provides a measure of the precision of the wannier function transformation . ] dimer ) linked to the bonding d@xmath58 and d@xmath59 orbitals occupied below the set of d@xmath60 , d@xmath61 , d@xmath62 orbitals ( in local coordinate system ) . top right panel : example of wannier function shared between fe - ga atoms . the red loop lies on an fe site . bottom panel : example of @xmath56-like wannier function centered on an fe site . fe atoms in brown , ga atoms in green . isosurfaces of different color correspond to opposite signs of the wf ] we conclude therefore that no simple fe charge state picture works for fega@xmath0 . the strong separation between occupied pseudocubic `` @xmath56 '' orbitals and empty `` @xmath63 '' orbitals is a distinguishing feature of the fe@xmath1 dimer , as is strong mixing with the ga @xmath64 orbitals . figure [ rho ] shows 3d isosurfaces of the charge density for the narrow fe @xmath2-dominated bands below and above the gap . consistent with the discussion provided above , the occupied states are a mixture of pseudocubic @xmath42 and @xmath43 charge ( @xmath28 along the dimer axis ) , with smaller contribution from @xmath41 , and is roughly circular in cross section around the dimer axis . the unoccupied states are a combination of pseudocubic @xmath63 states : @xmath45 orbitals directed along the dimer axis , and @xmath46 with lobes perpendicular to the dimer axis . this density is decidedly non - circular in cross section . both densities reflect antibonding character , or at least nonbonding character , with no evidence of bonding charge . there is negligible contribution to the density that is displayed from ga sites . the very small dispersion of the bands bordering the gap suggests rather localized molecular orbitals as the basic underlying feature for near - gap states . isosurface plots of states in fega@xmath0 . left panel : charge density obtained using xcrysden @xcite corresponding to the states at the bottom of the gap in fega@xmath0 ( energy range -0.3 - 0 ev ) . right panel : charge density of the states above the gap in fega@xmath0 ( energy range 0.6 - 0.9 ev ) . the strong directionality and the antibonding or nonbonding character of the charge along the fe - fe bond is evident . while the charge remains localized around fe atoms , the differing characters of the contributing states is obvious.,title="fig : " ] isosurface plots of states in fega@xmath0 . left panel : charge density obtained using xcrysden @xcite corresponding to the states at the bottom of the gap in fega@xmath0 ( energy range -0.3 - 0 ev ) . right panel : charge density of the states above the gap in fega@xmath0 ( energy range 0.6 - 0.9 ev ) . the strong directionality and the antibonding or nonbonding character of the charge along the fe - fe bond is evident . while the charge remains localized around fe atoms , the differing characters of the contributing states is obvious.,title="fig : " ] + we have performed calculations in which both @xmath4 ( ga ) and @xmath2 ( fe ) substitutional doping mechanisms have been explored in the 12-atom cell described above : either one ge or one zn were substituted on ga1 or ga2 sites ( doping level @xmath18= 0.25 , 0.5 , respectively ) and fe was substituted by either one co or one mn ( doping level @xmath65= 0.25 ) . within lda our results are consistent with those presented by singh within gga.@xcite due to the steep and large dos on either side of the gap , an itinerant fm state is obtained in all cases of low doping ( stoner mechanism for itinerant ferromagnetism ) . however , the magnetic state is different for the two types of directions of doping , as we will describe below . the scenario in which semiconducting fega@xmath0 has singlet fe@xmath1 dimers ( an antiferromagnetic state in a dft - based calculation ) was proposed by yin and pickett based on lda+@xmath22 calculations.@xcite as mentioned above , for the stoichiometric compound , lda and lda+@xmath22 ( @xmath22= 3 ev ) give very similar band gaps consistent with the experiment . when doping occurs , and especially when substitution is done on the fe site , effects of correlation are likely to become more evident . we also pursue this goal in the present section . table [ table1 ] shows the magnetic moments for each of the atoms in the metal dimers , the total magnetic moment in the cell , and the band gap for each of the cases studied in the corresponding magnetic ground state . this table will guide much of the discussion that follows . mn@xmath67ga@xmath0 within lda ( upper panel ) and lda+@xmath22 ( @xmath22= 3 ev ) ( lower panel ) . hole doping shifts the fermi level into the four - band complex with @xmath41 , @xmath42 , @xmath43 character for the minority spin channel giving rise to a half metallic ( hm ) result . the onsite coulomb repulsion causes orbital reoccupation with shifts in the band structure of 1 ev , and introduces @xmath63-like states in the gap for the minority spin channel . positive / negative values of the dos correspond to the spin - up / spin - dn channel . the fermi level is set at zero.,title="fig : " ] + hole doping has been explored on the fe site ( replacing one fe by mn ) and by substitution on the ga site by zn : ga1 substitution corresponds to a doping level @xmath18= 0.25 , and the zn atom has a pair of fe atoms at a distance of 2.36 . ga2 substitution corresponds to a doping level @xmath18= 0.5 , and the substituted zn has also a pair of fe atoms at a distance of 2.50 and one of the fe of the other dimer at 2.39 ( see fig . [ struct ] ) . in both cases , only an itinerant fm state can be obtained within lda . a half metallic ( hm ) fm state is obtained for fe@xmath68mn@xmath67ga@xmath0 with magnetic moments of 0.76 @xmath15 on the mn atom , 0.26 @xmath15 on the paired fe , and the other fe@xmath1 dimer remaining non magnetic ( see table [ table1 ] ) . the corresponding density of states is shown in the upper panel of fig . [ bs_femn ] . it exhibits a simple exchange splitting plus rigid - band behavior characteristic of an itinerant fm . the single hole difference leaves e@xmath69 lying within the four @xmath2 bands of the metallic minority spin channel , which are very strongly mn character the hole remains on the mn atom . magnetism arises from the @xmath56 orbitals , with larger contribution from the d@xmath61 orbital at the fermi level . the exchange splitting is about 0.25 ev . for fega@xmath70zn@xmath71 an itinerant fm state emerges with magnetism arising also from the @xmath43 orbital . the hole is localized on the fe atoms closer to the substituted zn ( see table [ table1 ] always showing higher magnetic moments on the fe atoms closer to the zn , with the other fe atoms having a significantly lower magnetic moment or even being nonmagnetic ) . experimental data on mn- and zn - doped fega@xmath0 differ from the lda - based predictions of emergence of a conducting fm state . @xmath24-type doping shifts the fermi level into the valence band and gives rise to metallic behavior , whereas transport measurements for hole doped fega@xmath0 show no insulator - metal transition , but instead semiconducting behavior , not consistent with itinerant behavior . for hole ( mn ) doping on the fe site , for the lower @xmath22 value of 2 ev , only fm ordering within the pairs can be obtained . a strong moment of 1.43@xmath15 on mn ( s=@xmath72 ) is obtained . the resulting electronic structure has the same features as that obtained within lda : a hm fm state . thus at @xmath22=2 ev correlation effects do not appear to be significant . for @xmath73 3 ev the picture changes qualitatively . not only can states with the magnetic moments being aligned and antialigned be obtained , but the afm alignment becomes energetically favored . the lower panel of fig . [ bs_femn ] shows the lda+@xmath22 ( @xmath22= 3 ev ) orbital resolved density of states of the mn atom in fe@xmath68mn@xmath67ga@xmath0 for the energetically favored afm alignment within the dimers . the magnetic moments are strengthened with @xmath22 ( see table [ table1 ] ) . while hm character remains , the onsite coulomb repulsion shifts certain @xmath36 bands by more than 1 ev giving rise to distinctive changes in configuration . a band with @xmath46 character appears within the gap in the minority spin channel , in contrast to the lda case . the complete @xmath46 spectral density within the gap strongly implies that for an isolated mn dopant ( _ i.e. _ the low doping limit ) an unoccupied @xmath46 gap state will appear , perhaps near mid - gap , when strong correlations are considered . the band shifts are even more noticeable for higher @xmath22 values , with the majority gap closing accompanied by an increase in the magnetic moments of mn and fe that reach values of 3.00@xmath15 and 2.05@xmath15 , respectively , and a widening of the bands around the fermi level . co@xmath67ga@xmath0 within lda ( upper panel ) and lda+@xmath22 ( lower panel ) . electron doping primarily drives the fermi level into the conduction band giving rise to a half metallic state upon doping . positive / negative values of the dos correspond to the spin - up / spin - dn channel . the fermi level is set at zero . ] electron doping at the fe site ( by co ) and at the ga site ( by ge ) also gives rise to an itinerant fm state within lda . in this case , magnetism arises from the @xmath63-like orbitals and the total moment of 1 @xmath15 is spread rather equally across all transition metal ions the electron is less localized ( see table [ table1 ] ) . there are only small differences in the dos for electron doping on the fe or ga site : low level electron doping with co for fe gives a simple rigid band shift downward of the majority spin channel ; when doping with ge different splittings appear for @xmath45 , @xmath46 orbitals due to the change in local environment with respect to the co - doped compound ( compare top panel of fig . [ dos_feco ] for fe@xmath68co@xmath67ga@xmath0 with fig . [ dos_ge ] for fega@xmath14ge@xmath14 ) . from experiments , co - doped fega@xmath0 behaves differently from the ge - doped counterpart . the ge - doped compound is ferromagnetic with a positive curie - weiss temperature whereas experiments for co - doping show an afm ordering at low t. dft - based calculations do not account for that difference , always favoring a ferromagnetic alignment of moments . the main difference with respect to hole doping is the more delocalized picture with the moments being equally distributed among all the metal atoms , in contrast to hole doped fega@xmath0 where there is a greater localization of the magnetic moments . ge@xmath14 obtained within lda . the upper two panels show the dos of fe1 and fe2 with partially occupied @xmath45 and @xmath46 orbitals . the lower panel shows the dos at fe3/fe4 sites . electron doping through ge substitution pushes the fermi level into the conduction band for both spin channels and induces magnetic moments on fe supported by exchange splitting . positive / negative values of the dos correspond to the spin - up / spin - dn channel . the fermi level is set at zero . ] for the lower @xmath22 value of 2 ev correlation effects are not significant in electron doped fega@xmath0 : a hm fm state is derived . at @xmath47 3 ev a ground state with antialigned moments is obtained . the lower panel of fig . [ dos_feco ] shows the orbital resolved dos of the co atom in fe@xmath68co@xmath67ga@xmath0 for the lda+@xmath22 ( @xmath22= 3 ev ) state . the picture upon electron doping appears somewhat simpler than for hole doping : the various orbital projected densities of states can be visualized as being connected adiabatically to their @xmath22=0 position ( upper panel ) with a hm state persisting . the outstanding difference is that the occupied @xmath45+@xmath46 weight above the gap converts to predominantly @xmath45 . this change corresponds to co introducing an occupied @xmath45 majority gap state into the gap at low doping . the co ion remains low - spin , whereas the magnetic moment on fe strengthens upon introducing @xmath22(see table [ table1 ] ) . for @xmath22 @xmath74 4 ev noticeable shifts in the @xmath36-bands take place , with the fe moment increasing up to 2.05 @xmath15 for @xmath22= 5 ev while co remains in its low spin state , with induced moment of 0.2@xmath15 irrespective of the @xmath22 value over the range we studied . as mentioned above , there is no unambiguous picture about the presence / absence of magnetism and/or correlations in fega@xmath0 . in agreement with previous works , the observed gap for undoped fega@xmath0 ( formed between t@xmath75 pseudocubic states at the top of the valence band and e@xmath76 states at the bottom of the conduction band ) can be reproduced without the inclusion of magnetism and correlations . this analysis is consistent with susceptibility measurements that indicate a diamagnetic state below rt@xcite that is also supported by the absence of local moments in fe moessbauer spectra . @xcite the antiferromagnetic spin singlet scenario derived from lda+@xmath22 calculations ( @xmath22= 3 ev ) is consistent with transport and thermodynamic experiments , combined with neutron powder diffraction data , interpreted as showing that fega@xmath0 is a correlated band insulator with a complex antiferromagnetic ordering.@xcite photoemission experiments agree with the correlated picture since the observed band dispersions are well reproduced by calculations within the lda+@xmath22 scheme ( @xmath22= 3 ev).@xcite the presence of magnetism in undoped fega@xmath0 is also consistent with muon spin rotation studies that show a spin polaron band that requires the existence of fe moments.@xcite electron or hole doping described without including correlations gives rise to an itinerant fm state ( stoner magnetism ) without pre - formed moments in the undoped compound . this type of result is consistent with experiments that show that fega@xmath16ge@xmath17 is a weak itinerant ferromagnet , but are difficult to reconcile with the role of strong correlations in the fm instability found beyond y@xmath20= 0.13 . @xcite this calculated metallic fm state also contrasts with experiments in co - doped fega@xmath0 that show itinerant antiferromagnetic behavior and suggest the existence of in - gap states at low doping.@xcite the correlated picture shows that co substitution for fe introduces a localized electronic gap state and gives rise to antialigned moments within the dimers , very different from the uncorrelated picture and supporting the experiments . experimental data on mn and zn doping also differ from the lda - based predictions of emergence of a hm fm state . _ et al__@xcite showed that hole doping does not give rise to an insulator - to - metal transition . they found instead that both substitution of zn onto the ga site , or mn on the fe site introduces states into the semiconducting gap that remain localized , suggesting the formation of small magnetic polarons . using neutron powder diffraction measurements and dmft , they establish that the complex magnetic order for fega@xmath0 above room temperature is almost unaffected by hole doping even though dynamical correlation effects become stronger . some of these observations are again in better agreement with the correlated picture that suggests that mn doping produces an in - gap hole state and an afm alignment within the dimers ( even though the calculated state within lda+@xmath22 is hm ) . to summarize , density functional theory methods have been applied to follow the evolution of the magnetic and electronic properties of fega@xmath0 with doping , where experiment shows that complex behavior including magnetism and quantum critical behavior arises . a specific interest here is to probe the difference for both electron and hole doping when doping is done on the magnetic fe ( dimer ) site or on the surely uncorrelated ga site . doping on either site introduces gap states or moves the chemical potential into bands on either side of the gap . using conventional dft for weakly correlated materials , the behavior upon substitution on the fe site ( disturbing the dimers ) or on the ga site ( doping off the dimers ) is similar : an itinerant ferromagnetic state emerges , as might be anticipated with this approach . however , there are clear differences between electron and hole doping : whereas electron doping gives rise to a more itinerant and delocalized picture ( with magnetic moments evenly distributed for all the fe atoms in the unit cell ) , upon hole doping there is a higher degree of localization of the magnetic moments moments develop on the dimer closest to the dopant atom . evidently the response to doping of states below the gap is quite different from that of states above the gap . when applying the correlated lda+u method , hole doping again is quite different from electron doping . co always maintains its low spin state , while mn is strongly magnetic . whether this difference reflects differences in the host states above and below the gap , or is due more to differences between mn and co , remains unclear . strong correlation effects lead to inter - dimer effects : doping on one dimer induces strong moments on the other ( `` undisturbed '' ) fe@xmath1 dimer . correlation effects are needed to be able to reconcile theoretical predictions with some of the rich behavior observed in doped fega@xmath0 . understanding the interplay between localized and itinerant magnetism ( including in - gap states at low doping ) will be crucial to explain the peculiar puzzles posed by this intermetallic compound . we acknowledge helpful communication on this topic with d. j. singh . was supported by doe grant de - fg02 - 04er46111 , y.q . and w. e. p. were supported by nsf grant dmr-1207622 . gap closings are of particular interest in these studies of doping . to better represent gap behavior , we have applied the lda+ mbj potential functional . for stoichiometric fega@xmath0 the band gap value obtained using mbj is slightly increased with respect to that obtained within lda . the dos obtained within mbj has the same features as that obtained within lda ( compare top panel of fig . [ mbj ] with fig . [ dos_lda ] ) . dos in local coordinates for each fe in the unit cell of fega@xmath0 ( top ) , mn in fe@xmath68mn@xmath67ga@xmath0 ( middle ) , and fe in fega@xmath77zn@xmath67 ( lower panel ) within mbj . half metallicity is broken in the doped cases ( in contrast to the lda predictions ) . positive / negative values of the dos correspond to the spin - up / spin - dn channel . the fermi level is set at zero . ] when electron doped , the experimental and lda - derived states are metallic and that type of state remains within mbj . from table [ table1 ] mbj favors a more localized picture with respect to lda or lda+@xmath22 . for ge - doped fega@xmath0 the magnetic moments are disproportionated within mbj , being higher in the fe pair ( fe1/fe2 ) closer to the substituted ge , unlike the lda result where the moments were evenly distributed among all fe atoms in the cell ) . for co - doped fega@xmath0 , there is a slight increase in the magnetic moment of the fe paired with the co whose magnetic moment remains unchanged within mbj , as do the magnetic moments of the fe atoms in the other dimer . when hole doped , the electronic structure obtained within mbj is quite different from that derived within lda and closer to the experimental results : the half metallicity is broken for both mechanisms of hole doping and antialigned magnetic moments within the dimers are favored . for fe@xmath68mn@xmath67 , the state resulting from mbj consists of a fm ordering within both the mn - fe and fe - fe pairs but with the moments having different sign for the two different dimers . the result is a zero - gap state versus the half metallic one derived using lda . the corresponding dos plot is shown in the central panel of fig . clearly , mbj shifts bands breaking the half metallicity , this is possible due to the shift of the d@xmath61-like orbital for the minority spin channel above the fermi level for mn and its paired fe . unlike the lda result , the mbj outcomes are consistent with experiments favoring a gap opening and an afm alignment of the moments in the two different dimers for mn - doped fega@xmath0 . the response obtained for zn - doped fega@xmath0 within mbj is also different from that of lda . again , the half metallicity is broken by mbj and a semiconducting state arises ( see bottom panel of fig . [ mbj ] for @xmath65= 0.25 ) with the concomitant increase in the magnetic moments that are antialigned as shown in table [ table1 ] . for both doping levels the gap opening is again possible due to the splitting of the d@xmath61-like orbital for the minority spin channel of fe atoms , being shifted above the fermi level . as commonly found in other applications , mbj tends to increase the values of the magnetic moments ( table [ table1]).@xcite the shifts in the band structure and the increase in the magnetic moments with respect to lda are consistent with changes in the charge inside the atomic spheres : increased for fe and accordingly reduced for ga atoms .

insulating fega@xmath0 poses peculiar puzzles beyond the occurrence of an electronic gap in an intermetallic compound . this fe - based material has a very distinctive structural characteristic with the fe atoms occurring in dimers . the insulating gap can be described comparably well in either the weakly correlated limit or the strongly correlated limit within density functional theory viewpoints , where the latter corresponds to singlet formation on the fe@xmath1 dimers . though most of the calculated occupied wannier functions are an admixture of fe @xmath2 and ga @xmath3 or @xmath4 states , there is a single bonding - type wannier function per spin centered on each fe@xmath1 dimer . density functional theory methods have been applied to follow the evolution of the magnetic properties and electronic spectrum with doping , where unusual behavior is observed experimentally . both electron and hole doping are considered , by ge and zn on the ga site , and by co and mn on the fe site , the latter introducing direct disturbance of the fe@xmath1 dimer . results from weakly and strongly correlated pictures are compared . regardless of the method , magnetism including itinerant phases appears readily with doping . the correlated picture suggests that in the low doping limit mn ( for fe ) produces an in - gap hole state , while co ( for fe ) introduces a localized electronic gap state .

the study of the nucleon / hyperon mean free paths in nuclei / nuclear matter is an important aspect for us to know about the nucleon / hyperon in - medium properties , because the nucleon / hyperon mean free paths relate with the nucleon - nucleon / hyperon - nucleon interactions in nuclear matter . in nuclear reactions , the nucleon / hyperon mean free path is a useful concept for summarizing a large number of experimental data , because the nucleon / hyperon mean free path in a nucleus can be introduced from the cross sections of the nucleon / hyperon - nucleus reactions , which can be obtained from experiment directly . since the work of bethe @xcite in 1940 , there has been considerable interest in the study of the mean free path of a proton or neutron in the nuclear medium @xcite . however , theoretical calculation on hyperon mean free path in nuclear medium has never been done due to the scarce experimental data on hyperon - nucleon / nucleus interactions . recently , some experimental progresses on hyperon increased the necessary for the study of the hyperon mean free path in nuclei . in this paper , we will firstly attempt to calculate the hyperon mean free paths in nuclei in the framework of rmf theory . the rmf theory have been used to study the nucleon mean free paths @xcite . with rmf method , the nucleon mean free path was found to be in reasonable agreement with experimental data . thus , in present work , we also study the @xmath0 and @xmath1-hyperon mean free paths in nuclei with rmf method . to calculate the hyperon mean free path in a nucleus , the essential point is to construct the energy - dependent hyperon - nucleus optical potentials . in this work , the real scalar and vector potentials are given by rmf . we attempt to obtain the hyperon imaginary potentials from those of nucleons @xmath8 and @xmath9 by the relations : @xmath10 and @xmath3 . in fact , cooper _ et al . _ also suggested to obtain the hyperon imaginary potentials from those of nucleons by means of multiplying a factor about ten years ago@xcite . the depth of the imaginary potential for @xmath1-@xmath11ca , @xmath12 mev at @xmath13=100 mev , is compatible with the prediction of batty _ _ et al.__@xcite , @xmath14 mev . we find that , in the central of the nuclei , the hyperon mean free path is about @xmath6 fm for @xmath0 , and about @xmath7 fm for @xmath1 , depending on the hyperon incident energy . the hyperon mean free path decreases with the increase of the hyperon incident energy . this paper is organized as follows . in the subsequent section we outline the hyperon - nucleon interactions in rmf . the details of the complex potentials are described in sec . [ complex ] . then the formulas about the hyperon mean free path are deduced in sec . the results and discussions are given in sec . [ result ] . finally a brief conclusion is given in sec.[con ] . within the framework of rmf theory@xcite , the effective lagrangian for hyperon and nucleons can be written as@xcite @xmath15 where @xmath16 is the standard lagrangian for nucleons @xmath17 with @xmath18 it involves nucleons @xmath19 , scalar @xmath20 mesons @xmath21 , vector @xmath22 mesons @xmath23 , vector isovector @xmath24 mesons @xmath25 , and the photon @xmath26 . the parameter of the nucleonic sector ( nl - sh ) are adopted from ref . @xcite , which describes properties of nuclear matter as well as of finite nuclei reasonably well . the nucleonic parameter set is presented in table [ nparameter ] . .the parametrization of the nucleonic sector ( nl - sh ) . the masses are given in ( mev ) and the coupling @xmath27 in ( @xmath28 ) . [ cols="^,^,^,^,^,^",options="header " , ] and @xmath29 , the lagrangian for hyperon yields @xmath30 where y stands for the hyperon @xmath0 and @xmath1 , and @xmath31 is the hyperon field . the pauli matrices for nucleons and hyperons are written as @xmath32 with @xmath33 being the third component . the ( iso)vector coupling constants for hyperon , @xmath34 and @xmath35 , are determined from the constituent quark model ( su(6 ) symmetry ) , namely : @xmath36 @xmath37 the scalar coupling constant for hyperon , @xmath38 , are determined by fitting to the potential depth of the corresponding hyperon in normal nuclear density with the following relation@xcite @xmath39 where @xmath40 is the hyperon potential depth in normal nuclear density , @xmath41 and @xmath42 are the values of @xmath20 and @xmath43 fields at saturation , and @xmath44=0.597 and @xmath45=0.146 @xmath46 for the set nl - sh . it is well known that the potential well depth of @xmath0 hyperon in nuclear matter is about @xmath47 mev , so we use @xmath48 mev to obtain the coupling constant @xmath49 . however , the experimental data on @xmath1 hypernuclei are very few . dover and gal@xcite analyzed old emulsion data on @xmath1 hypernuclei and concluded a nuclear potential well depth of @xmath50 to @xmath51 mev . fukuda et al . @xcite fitted the very - low - energy part of @xmath1 hypernuclear spectrum in the @xmath52c(k@xmath53,k@xmath54)x reaction and estimated the value of @xmath55 to be @xmath56 mev . recently , e885 at the ags @xcite have indicated a potential depth of @xmath57 mev or less . note that these @xmath1 potential depth data are estimated based on woods - saxon potentials . here , we choose @xmath58 mev to fix @xmath59 . according to the nl - sh parameter set for nucleons , we easily obtained the coupling constants @xmath60 , @xmath61 , @xmath62 and @xmath63 . for the nucleon / hyperon , within the framework of rmf , its dirac equation of motion is given by @xmath64\psi_\mathrm{b } % = \varepsilon = e\psi _ \mathrm{b},\ ] ] where @xmath65 and @xmath66 are the scalar potential and vector potential for the baryons . b stands for the baryons n , @xmath0 , @xmath1 , respectively . for simplicity , we neglect the coulomb interaction . to describe the absorptive properties of baryons in nuclei , we set @xmath65 and @xmath66 are the complex numbers in present work . here both of the complex potentials @xmath65 and @xmath66 are written as @xmath67 the real parts of the scalar potential and vector potential for baryons @xmath68 are @xmath69 if we neglected the contributions of isospin to the vector potential of the hyperons and nucleons , the real part of the scalar potential and vector potentials @xmath70 for hyperons can be related to those of nucleons with @xmath71 where @xmath72 and @xmath73 . from rmf theory , we can obtain the real parts of the complex potentials only . to calculate the hyperon mean free path , we have to know the imaginary part of complex potentials . since there is a simple linear relation between the real part of the hyperon scalar ( vector ) potential @xmath74 ( @xmath75 ) and the nucleon scalar ( vector ) potential @xmath76 ( @xmath77 ) ( see eqs.([sy ] ) and ( [ vy ] ) ) , the same linear relation should exist in the relation between the hyperon complex potential @xmath78 ( @xmath79 ) and the nucleon complex potential @xmath80 ( @xmath81 ) , which are @xmath82 according to the above suppose , the imaginary potentials of the hyperon - nucleus potentials can be derived from those of nucleons with the following relations : @xmath83 the imaginary parts of the nucleon - nucleus potentials @xmath8 and @xmath9 are extracted from ref . @xcite directly , which determined by fitting a large number of proton elastic scattering data . in fact , to estimate the imaginary parts of the hyperon - nucleus potentials , cooper _ et al . _ firstly assumed that the hyperon imaginary potentials could be obtained from those of nucleons by means of multiplying a factor about ten years ago@xcite . in rmf theory , in terms of the scalar potential @xmath84 and the vector potential @xmath85 , the momentum of a baryon propagating in the nuclei can be determined from @xmath86 replaced the complex potentials @xmath87 with eqs.([us ] , [ uv ] ) , the eq.([energymomentum ] ) can be rewritten as@xcite @xmath88 where @xmath89 with @xmath90 @xmath91 note that , @xmath92 can be identified as schrdinger equivalent potential which is the hyperon optical potential in the non - relativistic approach , eq.([energymomentum ] ) is identical to the non - relativistic dispersion relation , except the relativistic correction , @xmath93 . it is seen from the eqs.([potentialv ] , [ potentialw ] ) , the schrdinger equivalent potentials are energy dependent . at zero momentum of the hyperon , compared with the usual optical potential defined in rmf , @xmath94 , the schrdinger equivalent potentials have an additional term , which is the third term in eq.([potentialv ] ) and ( [ potentialw ] ) . how much the third term in eq.([potentialv])/ ( [ potentialw ] ) contributes to the schrdinger equivalent potentials , which we will discuss latter . with the imaginary part of hyperon - nucleus potentials given in eqs([ii ] , [ iii ] ) , we obtain the real part of the hyperon optical potential in eq . ( [ potentialv ] ) @xmath95 and the imaginary optical potential strength is then obtained in eq . ( [ potentialw ] ) : @xmath96 since the potentials @xmath97 and @xmath98 are complex , the hyperon momentum @xmath99 is also complex and can be expressed as @xmath100 the hyperon mean free path , @xmath101 , is related to the imaginary part of the hyperon momentum by @xmath102 from equations ( [ energymomentum2 ] ) , ( [ potential ] ) , ( [ momentum ] ) and ( [ lambda ] ) , it is easy to derive an analytical expression for @xmath101 , @xmath103^{\frac{1}{2 } } \bigg\}^{-\frac{1}{2}}. \label{lambda1}\end{aligned}\ ] ] for a hyperon propagating in a nucleus , its total energy @xmath104 is related to the incident energy @xmath13 by @xmath105^{1/2}},\end{aligned}\ ] ] where @xmath106 is the mass of nucleus . in fig . [ impp ] , we plot the imaginary parts of the nucleon - nucleus potentials as functions of radial radii @xmath107 at incident energy @xmath108 and 497.5 mev , respectively . obviously , the results of @xmath8 and @xmath9 in ref . @xcite are repeated . according to eqs . ( [ potentialvv ] ) , ( [ potentialww ] ) , we can easily obtain the energy dependent schrdinger equivalent potentials for the @xmath0 and @xmath1 hyperons . as an example , in fig . [ potentials ] , the nucleon , @xmath0 , @xmath1 schrdinger equivalent potentials in @xmath11ca are shown . to compare with the usual rmf optical potential , we set @xmath109 and plot the real schrdinger equivalent potential @xmath110 in fig . [ potentials ] ( b ) . because the imaginary nucleon - nucleus potentials are fitted at @xmath111 mev , we set @xmath112 mev and plot the imaginary part @xmath113 in fig . ( a ) . from the figure , we can see that the real parts of the depth of the schrdinger equivalent potential are @xmath114 , @xmath115 and @xmath116 mev for @xmath1 , @xmath0 and nucleon , respectively . the contributions of the third term in eqs . ( [ potentialvv ] ) to the real part of the schrdinger equivalent potential are about 7 mev for @xmath0-@xmath11ca and even reaches 35 mev for n-@xmath11ca , however the contributions to @xmath1-@xmath11ca is no more than 1 mev . as a whole , the depth of the usual rmf optical potential is deeper than the schrdinger equivalent potential . for n-@xmath11ca and @xmath0-@xmath11ca , there are large differences between the usual rmf optical potential and the schrdinger equivalent potential . there are few works about the imaginary optical potentials for the hyperons before . for @xmath1 , dover and gal suggested an imaginary part strength of @xmath117 mev@xcite . while batty _ _ suggested that the depth of imaginary optical potential for @xmath1 is about 3 mev , which is calculated with a @xmath118 potential with im@xmath119=0.04 fm @xcite . from the figure , we find our prediction for depth of the imaginary optical potential for @xmath1-ca at low incident energy ( @xmath120100 mev ) is @xmath121 mev , which is compatible with that of batty _ et al._. it indicates that our assumption for the imaginary potentials in eqs.([ii ] ) and ( [ iii ] ) is reasonable . for @xmath0 , we predicted the depth of the imaginary potential is @xmath5 7 mev . the depth of the imaginary potential for nucleon is @xmath122 11 mev . in the imaginary optical potential for @xmath123 and n , the deepest one is for n , the lowest one is for @xmath1 . the contributions of the last term in eq . ( [ potentialww ] ) to the imaginary potentials for @xmath123 and n is small , which are neglectable for @xmath123 . with the determined potentials , we can calculate the hyperon mean free path in nuclei . we show the hyperon @xmath124 mean free path in @xmath11ca and @xmath125pb as function of the radial distance @xmath107 in fig . [ zong ] . the solid , dotted curves correspond to the hyperon incident energies @xmath13 = 300 mev and 600 mev , respectively . in the center of the nuclei , the mean free path is about @xmath6 fm for @xmath0 , and about @xmath126 fm for @xmath1 , depending on the the hyperon incident energy . the mean free path is nearly independent of a certain nucleus . the hyperon mean free path increases rapidly at the surface of the nucleus due to the decrease of the hyperon - nucleus interactions . to see the effects of the hyperon incident energy on the mean free path clearly , we also plot the mean free path in @xmath11ca and @xmath125pb ( with @xmath1272 fm ) as a function of incident energy @xmath13 in fig . [ zong1 ] . from the figure , we find the hyperon mean free path decreases with the increase of the hyperon incident energy . at @xmath120 200 mev , the free mean path in @xmath11ca and @xmath125pb for @xmath0 ( @xmath1 ) is about 7.8 and 7 ( 4 and 3.5 ) fm , respectively , while at @xmath120 800 mev , which decreases to about 5 and 4.5 ( 2.5 and 2 ) fm . we have constructed the energy dependent schrdinger equivalent potentials for hyperon - nucleus , which are functions of scalar and vector potentials . the real scalar and vector potentials for hyperons are obtained from rmf theory directly . with the assumption , @xmath128 and @xmath129 , we relate the hyperon imaginary potentials with the nucleon s , which are @xmath130 and @xmath131 . the imaginary scalar and vector potentials , @xmath8 and @xmath9 , are obtained from ref . @xcite , which determined by fitting a large number of proton elastic scattering data . at low incident energy , the depth of the imaginary potential @xmath113 for hyperon - nucleon is on the order of several mev . we calculate the depth of the imaginary potential for @xmath1-@xmath11ca , @xmath12 mev at @xmath13=100 mev , which is compatible with the prediction of batty _ _ et al.__@xcite , @xmath132 mev . the imaginary potential is @xmath5 7 mev for @xmath0-@xmath11ca . with the determined potentials , we calculated the @xmath0 , @xmath1-hyperon mean free path in @xmath11ca , @xmath125pb with incident hyperon kinetic energy from 200 mev to 800 mev . in the interior of the nuclei , the hyperon mean free path less depends on a certain nucleus , and the mean free path is about @xmath6 fm for @xmath0 , and about @xmath7 fm for @xmath1 , depending on the hyperon incident energy . the hyperon mean free path decreases with the increase of the hyperon incident energy . because there were few works about the hyperon mean free path and hyperon - nucleus imaginary potential before , the present work is only an attempt in this field . many theoretical as well as experimental works are waiting to be done . besides , the study of the temperature dependence of the hyperon mean free path are also needed due to the appearance of hyperons in high - energy heavy - ion reactions . this work was supported by the natural science foundation of china ( 10275037 , 10575054 ) , the doctoral programme foundation of the china institution of higher education ( 20010055012 ) and the science and technology innovation foundation of nankai university , china . the authors would like to thank chunyan song for the useful discussion , and thank haixia an for our code . 99 h. a. bethe , phys . rev . * 57 * , 1125 ( 1940 ) . j. p. schiffer , nucl . phys . a**335 * * , 348 ( 1980 ) . m. t. collins and j. j. griffin , nucl . phys . a**348 * * , 63 ( 1980 ) . j. w. negele and k. yazaki , phys . rev . lett.*47 * , 71 ( 1981 ) .

the @xmath0- and @xmath1-hyperon mean free paths in nuclei are firstly calculated in the relativistic mean field ( rmf ) theory . the real parts of the optical potential are derived from the rmf approach , while the imaginary parts are obtained from those of nucleons with the relations : @xmath2 and @xmath3 . with the assumption , the depth of the imaginary potential for @xmath1 is @xmath4 3.5 mev , and for @xmath0 is @xmath5 7 mev at low incident energy . we find that , the hyperon mean free path decreases with the increase of the hyperon incident energies , from 200 mev to 800 mev ; and in the interior of the nuclei , the mean free path is about @xmath6 fm for @xmath0 , and about @xmath7 fm for @xmath1 , depending on the hyperon incident energy .

leptogenesis is an attractive way of generating the baryon number of the universe @xcite . the main idea put forward by fukugita and yanagida @xcite is that a lepton number asymmetry can be produced in the decays of heavy sterile neutrinos into leptons and higgs bosons and such an asymmetry is partially transferred to the baryon sector through electroweak sphaleron processes @xcite that violate @xmath2 and @xmath1 but not @xmath0 . the mechanism requires lepton number and @xmath5 violation , ensured by neutrino majorana masses and complex yukawa couplings . both masses and couplings form the well known seesaw model @xcite of light neutrino masses and thus establish a close relation between baryogenesis and low energy phenomenology . the third sakharov condition @xcite , the departure from thermal equilibrium , is satisfied in the delayed decay scenario . it means here that the majorana neutrinos live longer than the age of the universe when they approach non - relativistic temperatures and their production is boltzmann suppressed . it turns out that such condition puts an upper bound on the natural order of magnitude of the light neutrino masses @xcite well below the mass scale indicated by the atmospheric neutrino anomaly @xcite . this natural bound can be circumvented to some extent by fine - tuning the parameters to maximize the produced @xmath5 asymmetry but even so a quasi - degenerate light neutrino spectrum is excluded @xcite . as the seesaw provides a simple and elegant neutrino mass mechanism , it would be nice to have a naturally successful leptogenesis mechanism in the case where all heavy neutrinos have prompt rather than delayed decays . given that the heavy neutrino majorana masses and yukawa couplings provide the necessary @xmath0 and @xmath5 violations the main concern is to satisfy the out - of - equilibrium condition . this is naturally achieved by assuming that the heavy sterile neutrinos do not couple to the inflaton field and are only thermally produced during the radiation era from standard lepton and higgs particles . before the heavy neutrinos reach thermal equilibrium abundances a non - zero @xmath0 asymmetry can in principle be generated . that is the subject of this paper . in the next section we describe the mechanism in more detail and in the following ones we present the main results concerning first , the generation of the primary @xmath0 asymmetry and second , the protection mechanism that prevents @xmath0 from being completely washed out when the lepton number violating reactions reach thermal equilibrium . in the last section we summarize the mechanism and make a brief discussion . the seesaw mechanism adds to the standard model singlet ( left - handed ) neutrinos , @xmath6 , with heavy majorana masses and yukawa couplings to the standard lepton and higgs doublets , @xmath7 , @xmath8 , of the form @xmath9 spontaneous breaking of su(2 ) @xmath10 u(1 ) yields the light neutrino mass matrix ( @xmath11 ) @xmath12 the proper decay rate of @xmath6 into leptons and higgs is @xmath13 . delayed decay occurs when the ratio to the hubble expansion rate , @xmath14 , is small at the temperature @xmath15 . in the radiation era , @xmath16 , where @xmath17 denotes the number of relativistic degrees of freedom , 107.5 in the standard model . it is enough to compare the sum @xmath18 v^2}{10^{-3 } \ ; \mathrm{ev}}\ ] ] with the light neutrino scale @xmath19 $ ] to conclude that the delayed decay condition , @xmath20 , is in general in conflict with the atmospheric neutrino mass gap @xcite , @xmath21 , which implies @xmath22 . strictly speaking , the delayed decay scenario only requires that the lightest of the heavy neutrinos satisfies @xmath20 , and larger @xmath23 values are tolerated for a certain choice of parameters that maximize the @xmath5 asymmetry @xcite . but that is not the most natural picture , in particular if the light neutrinos are quasi - degenerate @xcite . as it turns out , there is a simple modification of the original leptogenesis mechanism capable of realizing the out - of - equilibrium condition and producing a large lepton asymmetry . according to the inflation paradigm @xcite , the radiation era starts when a scalar field , so - called inflaton , decays and transforms its energy into the form of particles so constituting the initial thermal bath . if sterile neutrinos do not couple to the inflaton , they can be only thermally produced from standard lepton and higgs particles through yukawa interactions . these processes are naturally out - of - equilibrium at very high temperatures and it takes some time before sterile neutrinos reach thermal equilibrium densities . during that period there is a clear departure from equilibrium and a @xmath0 asymmetry can be generated . in this scenario a primary lepton asymmetry is generated in the lepton doublet and sterile neutrino sectors . a fraction of it is transferred to the bottom @xmath24 , tau @xmath25 , charm @xmath26 and other right - handed fermions through slow yukawa interactions . another fraction is converted into baryon number by weak sphaleron processes . it is crucial for the survival of such baryon number and @xmath24 , @xmath25 , @xmath26 , ... asymmetries that the heavy majorana neutrinos decay and vanish from the universe before the weak sphalerons or @xmath24 , @xmath25 ( @xmath26 ) yukawa couplings enter in equilibrium , which happens at a temperature around @xcite @xmath27 ( @xmath28 ) . by that time the @xmath29 reactions mediated by off - shell majorana neutrinos should also be out - of - equilibrium . in other words , the moment when the @xmath29 reactions freeze out determines which decoupled sectors or quantum numbers protect and store a fraction of the primary @xmath0 asymmetry . their rate relates to the sum of light neutrino square masses @xmath30 $ ] as @xmath31 @xcite . they are out of equilibrium , i.e. , @xmath32 , at temperatures below @xmath33 we assume of course that the masses @xmath34 are larger than @xmath35 so that the heavy neutrinos vanish above that temperature . @xmath36 ( @xmath28 ) for neutrino masses @xmath37 ( @xmath38 ) , well above the atmospheric neutrino mass gap . as will be shown below , if @xmath39 then the baryon number sector and the right - handed quark @xmath24 have the dominant role in protecting @xmath0 and producing the present baryon number of the universe . if @xmath3 exceeds @xmath4 that role would be played by the charm quark @xmath26 but not efficiently enough to produce the present baryon asymmetry . we assume that the universe is initially empty of sterile neutrinos , not produced in the inflaton decay but only thermally from standard leptons and higgs . for definiteness we assume a reheating temperature much higher than the neutrino masses . the dominant production reaction is the higgs boson decay into leptons and neutrinos , @xmath40 , allowed by a higgs thermal mass , @xmath41 @xcite , much higher than the lepton thermal masses . the yukawa top quark scattering processes like @xmath42 and other channels contribute to not more than 40 % of the total @xmath6 production rate . the higgs boson proper decay rate into @xmath6 is @xmath43 per isospin state . in the radiation era the initially zero neutrino densities @xmath44 converge to the fermion equilibrium density @xmath45 with relaxation temperatures @xmath46 : @xmath47 as a consequence of the prompt decay assumption , @xmath48 , sterile neutrinos reach thermal equilibrium while they are still ultra relativistic . leptogenesis is dominated by the following @xmath5 asymmetric reactions : higgs decays into leptons and sterile neutrinos , inverse decays and scatterings of leptons off neutrinos . the asymmetries result from the interference between the diagrams on the left side of figure [ fig1 ] with the absorptive part of the respective diagram on the right - hand side . unitarity and @xmath49 invariance impose constraints on the reaction rates such that when all particles are in thermal equilibrium the asymmetry source terms cancel each other @xcite . generically denoting the difference between the total rates of a reaction @xmath50 and its @xmath5-conjugate @xmath51 as @xmath52 ( in a fixed comoving volume ) , the relevant constraints are expressed in terms of thermal equilibrium rates as : ( | n_a l_i ) _ + ( n_a l_i | ) _ = 0 , [ cpta ] + ( n_a l_i n_b l_j ) _ + ( n_b l_j n_a l_i ) _ = 0 , [ cptb ] + ( n_a l_i | ) _ + _ bj ( n_a l_i n_b l_j ) _ = 0 [ cptc ] . the evolution of the set of observables @xmath53 is governed by the boltzmann equations , whose general structure is , integrating over a constant comoving volume , @xmath54 . the first are transport terms and @xmath55 stand for source terms . the particle asymmetry sources vanish in thermal equilibrium but , as long as the @xmath6 neutrinos stay rarefied , scatterings and inverse decays do not match higgs decays and particle asymmetries develop in the various lepton flavors , sterile neutrinos , and higgs boson as well , as enforced by hypercharge conservation . the calculation of the source terms is simplified by employing boltzmann rather than quantum statistics as is common practice . relating the out - of - equilibrium and thermal equilibrium reaction rates as @xmath56 for @xmath57 , and @xmath58 for @xmath59 , one derives from the above @xmath49 and unitarity constraints the @xmath7 lepton number sources , @xmath60 where @xmath61 , @xmath62 and @xmath63 is the higgs abundance per isospin state . sterile neutrinos also have lepton number , @xmath64 for left - handed @xmath6 fields and @xmath65 for the conjugate @xmath66 . but what matters for particle propagation and transport of lepton number are the spin eigenstates rather than chiral states . when neutrinos are ultra relativistic , helicity and chiral states are almost identical , but when they are non - relativistic majorana masses take over and lepton number vanishes . in fact , a positive ( negative ) helicity eigenstate has a well defined lepton number expectation value equal to the neutrino speed @xmath67 ( @xmath68 ) . helicity is not an invariant quantity , however , there is a privileged reference frame , the comoving thermal bath rest frame . in that frame , isotropy ensures that the spin density matrix is diagonal in the helicity basis . this is not true in any other frame , and means that neutrinos can be divided in two populations of opposite helicities . under the non - equilibrium conditions of the leptogenesis era the two neutrino helicity populations develop unequal abundances and net lepton numbers in much the same way as the standard leptons . from the same decay , inverse decay and scattering processes one derives the @xmath69 lepton number source terms , @xmath70 where @xmath71 is a coefficient estimated to be of the order but larger than unit . summing both lepton and neutrino contributions , one obtains the total @xmath0 asymmetry as a function of the temperature : @xmath72 ^ 2}{(h^\dagger h)_{aa } } \frac{m_a m_b } { t_a^{2 } } f\left(\frac{t_a}{t } \right ) , \ ] ] where @xmath73 , @xmath74 is the massless fermion thermal abundance and @xmath75 . the other question is how fast majorana masses erase lepton number . in a decay @xmath40 , lepton number suffers a variation @xmath76 if @xmath6 has positive helicity , which occurs with a branching fraction of about @xmath77 . it makes an average variation @xmath78 per decay and the @xmath6 production rates through higgs decays are equal to @xmath79 with @xmath80 . thus , lepton number violation enters in equilibrium at a temperature @xmath81 , i.e. , @xmath82 the @xmath76 scattering processes mediated by majorana neutrinos like @xmath83 , that are so important when the majorana neutrinos are non relativistic and even at lower temperatures when they are off - shell , are completely negligible in comparison with the higgs boson decays @xmath84 transitions at relativistic temperatures . the relative strength is of the order of @xmath85 . for constants @xmath23 as large as 200 , the sterile neutrinos enter in thermal equilibrium at temperatures @xmath86 but the majorana masses , i.e. , the @xmath87 transitions only enter in equilibrium at @xmath88 , where @xmath89 denotes the largest of the majorana masses . @xmath90 is quite smaller than the relaxation temperatures @xmath91 and the larger the @xmath23 constants are the smaller is @xmath90 in comparison with @xmath91 . at temperatures above @xmath90 , @xmath0 is not significantly damped and is slowly transferred via yukawa interactions to the right - handed leptons and quarks , @xmath92 , @xmath93 and @xmath94 ( with the exception of the top , @xmath95 , very early in chemical equilibrium ) . the dominant processes are higgs boson decays into leptons @xcite , @xmath96 , kinematically forbidden in the case of quarks due to comparable quark thermal masses @xcite , and scatterings like @xmath97 , @xmath98 . these processes generate the otherwise zero asymmetries @xmath99 and @xmath100 , @xmath101 . for right - handed quarks , the contrast asymmetries with respect to the up quark @xmath102 are the appropriate charges because , contrary to the individual flavor asymmetries , they are conserved by qcd instantons @xcite and are only violated by yukawa couplings . their growth rates are determined by the total yukawa production rates , @xmath103 , of @xmath104 , @xmath105 , @xmath94 , and degeneracy parameters @xmath106 as , _ e_i = 2 ( _ l_i - _ - _ e_i ) _ e_i y_f , [ qei ] + _ u_i = 2 ( _ q_i + _ - _ u_i ) _ u_i y_f , [ qui ] + _ d_i = 2 ( _ q_i - _ - _ d_i ) _ d_i y_f , [ qdi ] neglecting flavor violation and @xmath102 yukawa coupling . @xmath107 scale as the temperature and the ratios @xmath108 define equilibrium temperatures , @xmath109 , under which the respective detailed balance equations are enforced . we estimate @xmath110 and @xmath111 . on the other hand , weak sphalerons transfer part of the @xmath0 asymmetry to the baryon sector and the baryon number growth rate @xcite in a volume @xmath112 is given by @xmath113 where @xmath114 is the sphaleron diffusion rate . the integration of the above equations requires knowing the relations between the chemical potentials and the leptogenesis sources given in equations ( [ lis ] ) and ( [ lbs ] ) . in general there is no direct relation because the partial particle asymmetries are not conserved due to flavor changing yukawa couplings . as soon as the first neutrino , say @xmath115 , is in equilibrium ( @xmath116 ) , the heavy neutrino and standard lepton flavors are rapidly violated enforcing their chemical potentials @xmath117 to be identical : @xmath118 . the higgs communicates also its asymmetry to @xmath95 and top quark doublet : @xmath119 . because the @xmath120 asymmetries increase rapidly with time , initially as @xmath121 , we concentrate on the latest period before @xmath90 , where sterile neutrinos are already in equilibrium but majorana masses do not violate lepton number fast enough to yield null neutrino chemical potentials . equations ( [ qei])-([dbdt ] ) can be written in the form @xmath122 where the coefficients @xmath123 are constant in a radiation dominated universe but depend on the all set of chemical potential constraints . assuming that @xmath90 lies above @xmath124 one obtains for @xmath125 and the baryon number , @xmath126 and @xmath127 , respectively , while @xmath128 because @xmath129 . the integration in time gives asymmetries per unity of entropy equal to @xmath130 ^ 2}{[(h^\dagger h)_{aa}]^2 } \frac{m_a m_b}{t_a t } g\left ( \frac{t_a}{t } \right ) \ ; , \ ] ] where @xmath131 . the @xmath120 asymmetries grow with @xmath132 ( @xmath133 ) down to the temperature @xmath134 where lepton number starts to be rapidly violated . then , a new constraint is enforced , @xmath135 , that would cause @xmath0 and all the other asymmetries to vanish if there were no decoupled @xmath120 asymmetries . @xmath0 falls down to the level of the dominant right - handed fermion asymmetries , or total baryon number , and is constrained to be a linear combination of them . they remain marginally damped as long as the respective yukawa processes or weak sphaleron processes are slow , as implied by equations ( [ qei])-([dbdt ] ) . as the temperature drops down a new chemical potential constraint is enforced whenever any of these interactions enters in equilibrium and @xmath0 falls down further becoming a linear combination of the remanescent set of decoupled asymmetries . the final @xmath0 asymmetry is determined by the @xmath120 asymmetries that are still approximately conserved when the @xmath0 violating , @xmath84 , processes become negligible . from then on @xmath0 is exactly conserved . assuming that this happens above @xmath136 , the @xmath24 yukawa and weak sphalerons are still out of equilibrium and the final @xmath0 asymmetry is dominated by the @xmath125 and @xmath2 contributions : @xmath137 the stars mean that @xmath138 , @xmath139 and @xmath140 are calculated from equation ( [ qchi ] ) at the temperature @xmath141 and remain practically constant after that . the final @xmath0 asymmetry is @xmath142 ^ 2}{[(h^\dagger h)_{aa}]^2 } \frac{m_a m_b}{t_a t_\mathrm{in } } g\left ( \frac{t_a}{t_\mathrm{in } } \right ) \ ; .\ ] ] as the temperature goes down and all right - handed leptons and quarks gradually enter in equilibrium @xmath0 is redistributed among particles to eventually yield the present baryon number asymmetry @xmath143 @xcite . one obtains @xmath0 and baryon asymmetries at exactly the observed order of magnitude , @xmath144 per unit of entropy , taking precisely the prompt decay ratio values indicated by the atmospheric neutrino evidence , @xmath145 , without assuming any particular hierarchy between yukawa couplings . in the solution described above @xmath0 is mainly originated from the baryon number and @xmath24 asymmetry @xmath146 . it assumes that @xmath76 reactions go out of equilibrium before the @xmath24 and weak sphaleron processes enter in equilibrium @xcite at @xmath147 . as equation ( [ tout ] ) indicates this requires that the light neutrino mass scale @xmath3 is not larger than @xmath4 , which is still one order of magnitude higher than the atmospheric neutrino mass gap . if @xmath3 goes beyond @xmath4 , the charm @xmath26 replaces @xmath24 in the role of protecting @xmath0 but @xmath148 is too small unless it is compensated by a strong hierarchy in the yukawa coupling structure . to conclude , leptogenesis can operate in a regime where all sterile neutrinos , including the lightest one , decay promptly when they vanish from the universe . the light neutrino mass scale @xmath149 determines the natural order of magnitude of the final baryon asymmetry . first , @xmath3 values between the atmospheric neutrino mass gap and @xmath4 imply that the heavy neutrinos decay promptly with ratios @xmath150 , and reach thermal equilibrium at temperatures 10 times or more above their masses . during thermal production , @xmath0 is generated in both lepton and sterile neutrino sectors reaching the @xmath151 level . later , when the majorana masses rapidly violate lepton number , @xmath0 is effectively protected by the chemically decoupled @xmath24 quark and baryon number sector if , again , @xmath39 , provided that the majorana masses lie above the @xmath24 and weak sphaleron equilibrium temperature around @xmath136 . then , @xmath0 falls down 3 orders of magnitude , which meets exactly the observed baryon number asymmetry @xmath2 without any further conditions . the dependence of @xmath0 on the parameters of the theory is totally different from the delayed decay scenario . in the usual @xmath152 hierarchical case , the lepton asymmetry generated in @xmath153 decays vanishes in the limit of degenerate light neutrinos and @xcite is essentially proportional to the atmospheric neutrino mass gap and the imaginary part @xmath154 $ ] , where @xmath155 is an element of the orthogonal complex matrix @xmath156 that parametrizes the couplings @xmath157 , and @xmath158 is the diagonalized neutrino mass matrix . on the contrary , in the thermal production regime @xmath0 does not vanish in the degenerate limit and goes with @xmath159 $ ] . equations ( [ qchi ] ) and ( [ bmlstar ] ) yield @xmath24 and @xmath0 asymmetries of the required @xmath160 order for a wide range of parameters . they only depend on light and heavy neutrino mass ratios and matrix @xmath161 . m. gell - mann , p. ramond , r. slansky . in : _ supergravity , proceedings of the workshop , stony brook , n. y. , 1979 _ , ed . by p. van nieuwenhuizen , d. freedman ( north - holland , amsterdam , 1979 ) t. yanagida . in : _ proceedings of the workshop on unified theories and baryon number in the universe , tsukuda , japan , 1979 _ , ed . by a. sawada , a. sugamoto ( kek report no . 79 - 18 , tsukuda , 1979 ) t. yanagida : prog . th . phys . b * 135 * , 66 ( 1979 )

leptogenesis is studied within the seesaw neutrino mass model in a regime where all sterile neutrinos have prompt rather than delayed decays . it is shown that during neutrino thermal production lepton asymmetries are generated in both active lepton and sterile neutrino sectors . the large @xmath0 asymmetry is slowly pumped into the chemically decoupled right - handed quarks and leptons and baryon number sector which later protect @xmath0 from fast @xmath1 violating processes . the dependence of the final baryon asymmetry on couplings and masses is totally different from the decay scenario . @xmath2 does not vanish in the limit of degenerate light neutrinos and the observed asymmetry is naturally obtained for a sum of square masses , @xmath3 , between the atmospheric neutrino mass gap and @xmath4 .

the concept of quantitative operator @xmath0-theory was first introduced in @xcite and was set - up in full generality for filtered @xmath1-algebras in @xcite . the aim of this paper is to develop techniques of quantitative operator @xmath0-theory to compute @xmath0-theory of @xmath1-algebras . in particular , we introduce a concept of finite asymptotic nuclear decomposition for filtered @xmath1-algebras . this @xmath2-algebraic concept can be viewed as the noncommutative analogue of metric spaces with finite asymptotic dimension . we establish an algorithm for computing @xmath0-theory of @xmath1-algebras with finite asymptotic nuclear decomposition . as a consequence , we prove the knneth formula for @xmath1-algebras in this class . the key idea is that the finite asymptotic nuclear decomposition for a @xmath1-algebra would allow us to compute quantitative @xmath0-theory for the @xmath1-algebra at each scale and the @xmath0-theory of the @xmath1-algebra can then be computed by letting the scale go to infinity . the main technical tool to compute quantitative @xmath0-theory is controlled mayer - vietoris pairs and attached controlled six term exact sequences . the paper is organized as follows . in section 1 , we give from @xcite an overview of quantitative @xmath0-theory . in section 2 we introduce the concept of a controlled mayer - vietoris pair . this is the key ingredient to define later on the class of @xmath1-algebras with finite asymptotic nuclear decomposition . typical example of these objects arise from roe algebras and more generally from @xmath2-algebras of tale groupods . in section @xmath3 is stated for a controlled mayer - vietoris pair the controlled six term exact sequence . we apply this sequence to @xmath0-contractibility of @xmath2-algebra . section 4 is devoted to the quantitative knneth formula , which implies the knneth formula in @xmath0-theory . we show that examples of filtered @xmath2-algebras for which the quantitative knneth formula holds are provided by crossed product of @xmath2-algebras by finitely generated groups satisfying the baum - connes conjecture with coefficients . the main result of these section in that the quantitative knneth formula is stable under decomposition by controlled mayer - vietoris pair . in section 5 we introduce the class of @xmath1-algebras with finite asymptotic nuclear decomposition . this class is the class of filtered @xmath2-algebras @xmath4 for which there exists a integer @xmath5 such that at every order @xmath6 , the @xmath2-algebra @xmath4 decomposes in @xmath5 steps under controlled mayer - vietoris into @xmath2-algebras in the bootstrap category . we then provide an algorithm for computing @xmath0-theory for @xmath2-algebras in this class . in particular we show that these @xmath2-algebras satisfy the quantitative knneth formula . as a consequence , we show that the uniform roe algebra of a discrete metric space with bounded geometry and with finite asymptotic dimension satisfies the kunneth formula . in this section , we recall the basic concepts of quantitative @xmath0-theory for filtered @xmath1-algebras and collect the main results of @xcite concerning quantitative @xmath0-theory that we shall use throughout this paper . roughly speaking , quantitative @xmath0-theory is the abelian groups of @xmath0-theory elements at each given scale and @xmath0-theory can be obtained as an inductive limit of quantitative @xmath0-groups ( see corollary [ cor - limit - quant - k - groups ] ) . the key point is that quantitative @xmath0-theory is in numerous geometric situations more computable that usual @xmath0-theory . the structure of filtered @xmath2-algebras allows us to talk about the scale of elements in the @xmath1-algebras . a filtered @xmath2-algebra @xmath4 is a @xmath2-algebra equipped with a family @xmath7 of closed linear subspaces indexed by positive numbers such that : * @xmath8 if @xmath9 ; * @xmath10 is stable by involution ; * @xmath11 ; * the subalgebra @xmath12 is dense in @xmath4 . if @xmath4 is unital , we also require that the identity @xmath13 is an element of @xmath14 for every positive number @xmath6 . the elements of @xmath10 are said to have * propagation @xmath6*. many examples of filtered @xmath2-algebras arise from geometry . typical examples are provided by roe algebras , group and crossed - product algebra @xcite , groupoid algebras ( see section [ subsection - groupoid ] ) and finitely generated @xmath2-algebras . indeed all filtered @xmath2-algebras are associated with a * length function * : let @xmath4 be a @xmath2-algebra and assume that there exists a function @xmath15 such that * @xmath16 ; * @xmath17 for all @xmath18 and @xmath19 in @xmath4 ; * @xmath20 for all @xmath18 in @xmath4 ; * @xmath21 for all @xmath18 in @xmath4 and @xmath22 in @xmath23 ; * @xmath24 for all @xmath18 and @xmath19 in @xmath4 ; * @xmath25 is closed in @xmath4 for all positive number @xmath6 ; * @xmath26 is dense in @xmath4 . if we set @xmath27 , then @xmath4 is filtered by @xmath7 . it is straightforward to show that the category of filtered @xmath2-algebras is equivalent to the category of @xmath2-algebras equipped with a length function . in essence , we study geometric @xmath2-algebras just as group theorists study geometric group theory . let @xmath4 and @xmath28 be respectively @xmath2-algebras filtered by @xmath7 and @xmath29 . a homomorphism of @xmath2 -algebras @xmath30 is a * filtered homomorphism * ( or a * homomorphism of filtered @xmath2-algebras * ) if @xmath31 for any positive number @xmath6 . if @xmath4 is not unital , let us denote by @xmath32 its unitarization , i.e. , @xmath33 with the product @xmath34 for all @xmath35 and @xmath36 in @xmath32 . then @xmath37 is filtered with @xmath38 we also define @xmath39 . let @xmath4 be a unital filtered @xmath2-algebra . for any positive numbers @xmath6 and @xmath40 with @xmath41 , we call * an element @xmath42 in @xmath4 an @xmath40-@xmath6-unitary if @xmath42 belongs to @xmath10 , @xmath43 and @xmath44 . the set of @xmath40-@xmath6-unitaries on @xmath4 will be denoted by @xmath45 . * an element @xmath46 in @xmath4 an @xmath40-@xmath6-projection if @xmath46 belongs to @xmath10 , @xmath47 and @xmath48 . the set of @xmath40-@xmath6-projections on @xmath4 will be denoted by @xmath49 . then @xmath40 is the called the control and @xmath6 is called the propagation of the @xmath40-@xmath6-projection or of the @xmath40-@xmath6-unitary . notice that an @xmath40-@xmath6-unitary is invertible , and that if @xmath46 is an @xmath40-@xmath6-projection in @xmath4 , then it has a spectral gap around @xmath50 and then gives rise by functional calculus to a projection @xmath51 in @xmath4 such that @xmath52 . recall the following from ( * ? ? ? * lemma 1.7 ) the following result that will be used quite extensively throughout the paper . [ lemma - almost - closed ] let @xmath4 be a @xmath2-algebra filtered by @xmath7 . 1 . if @xmath46 is an @xmath40-@xmath6-projection in @xmath4 and @xmath53 is a self - adjoint element of @xmath10 such that @xmath54 , then @xmath53 is an @xmath40-@xmath6-projection . in particular , if @xmath46 is an @xmath40-@xmath6-projection in @xmath4 and if @xmath53 is a self - adjoint element in @xmath10 such that @xmath55 , then @xmath53 is a @xmath56-@xmath6-projection in @xmath4 and @xmath46 and @xmath53 are connected by a homotopy of @xmath56-@xmath6-projections . if @xmath4 is unital and if @xmath42 is an @xmath40-@xmath6-unitary and @xmath57 is an element of @xmath10 such that @xmath58 , then @xmath57 is an @xmath40-@xmath6-unitary . in particular , if @xmath42 is an @xmath40-@xmath6-unitary and @xmath57 is an element of @xmath10 such that @xmath59 , then @xmath57 is an @xmath60-@xmath6-unitary in @xmath4 and @xmath42 and @xmath57 are connected by a homotopy of @xmath60-@xmath6-unitaries . if @xmath46 is a projection in @xmath4 and @xmath53 is a self - adjoint element of @xmath10 such that @xmath61 , then @xmath53 is an @xmath40-@xmath6-projection . if @xmath4 is unital and if @xmath42 is a unitary in @xmath4 and @xmath57 is an element of @xmath10 such that @xmath62 , then @xmath57 is an @xmath40-@xmath6-unitary . let us also mention the following result concerning homotopy up to stabilization of products of @xmath40-@xmath6-unitaries ( * ? ? ? * corollary 1.8 ) . [ cor - example - homotopy ] let @xmath40 and @xmath6 be positive numbers with @xmath63 and let @xmath4 be a unital filtered @xmath2-algebra . 1 . let @xmath42 and @xmath57 be @xmath40-@xmath6-unitaries in @xmath4 , then @xmath64 and @xmath65 are homotopic as @xmath66-@xmath67-unitaries in @xmath68 ; 2 . let @xmath42 be an @xmath40-@xmath6-unitary in @xmath4 , then @xmath69 and @xmath70 are homotopic as @xmath66-@xmath67-unitaries in @xmath68 . for purpose of rescalling the control and the propagation of an @xmath40-@xmath6-projection or of an @xmath40-@xmath6-unitary , we introduce the following concept of control pair . a control pair is a pair @xmath71 , where * @xmath72 is a positive number with @xmath73 ; * @xmath74 is a map such that there exists a non - increasing map @xmath75 , with @xmath76 . the set of control pairs is equipped with a partial order : @xmath77 if @xmath78 and @xmath79 for all @xmath40 in @xmath80 . recall the following from ( * ? ? ? * corollary 1.31 ) . [ prop - conjugate ] there exists a control pair @xmath81 such that the following holds : for any unital filtered @xmath2-algebra @xmath4 , any positive numbers @xmath40 and @xmath6 with @xmath82 and any homotopic @xmath40-@xmath6-projections @xmath83 and @xmath84 in @xmath85 , then there is for some integers @xmath86 and @xmath87 an @xmath88-@xmath89-unitary @xmath90 in @xmath91 such that @xmath92 for any @xmath5 integer , we set @xmath93 and @xmath94 . for any unital filtered @xmath2-algebra @xmath4 , any positive numbers @xmath40 and @xmath6 and any positive integer @xmath5 , we consider inclusions @xmath95 and @xmath96 this allows us to define @xmath97 and @xmath98 for a unital filtered @xmath2-algebra @xmath4 , we define the following equivalence relations on @xmath99 and on @xmath100 : * if @xmath46 and @xmath53 are elements of @xmath101 , @xmath87 and @xmath102 are positive integers , @xmath103 if there exists a positive integer @xmath86 and an element @xmath104 of @xmath105)$ ] such that @xmath106 and @xmath107 . * if @xmath42 and @xmath57 are elements of @xmath100 , @xmath108 if there exists an element @xmath104 of @xmath109)$ ] such that @xmath110 and @xmath111 . if @xmath46 is an element of @xmath101 and @xmath87 is an integer , we denote by @xmath112_{{\varepsilon},r}$ ] the equivalence class of @xmath113 modulo @xmath114 and if @xmath42 is an element of @xmath100 we denote by @xmath115_{{\varepsilon},r}$ ] its equivalence class modulo @xmath114 . let @xmath6 and @xmath40 be positive numbers with @xmath41 . we define : 1 . @xmath116 for @xmath4 unital and @xmath117_{{\varepsilon},r}\in { \operatorname{p}}^{{\varepsilon},r}({\tilde{a}})\times{\mathbb n}/\sim { \text { such that } } { \operatorname{rank}}\kappa_0(\rho_{a}(p))=l\}\ ] ] for @xmath4 non unital ( @xmath118 being the spectral projection associated to @xmath119 ) ; 2 . @xmath120 , with @xmath121 if @xmath4 is already unital . then @xmath122 turns to be an abelian group ( * ? ? ? * lemma 1.15 ) , where @xmath123_{{\varepsilon},r}+[p',l']_{{\varepsilon},r}=[{\operatorname{diag}}(p , p'),l+l']_{{\varepsilon},r}\ ] ] for any @xmath112_{{\varepsilon},r}$ ] and @xmath124_{{\varepsilon},r}$ ] in @xmath122 . according to corollary [ cor - example - homotopy ] , @xmath125 is equipped with a structure of abelian group such that @xmath126_{{\varepsilon},r}+[u']_{{\varepsilon},r}=[{\operatorname{diag}}(u , v)]_{{\varepsilon},r},\ ] ] for any @xmath115_{{\varepsilon},r}$ ] and @xmath127_{{\varepsilon},r}$ ] in @xmath125 . recall from ( * ? ? ? * corollaries 1.19 and 1.21 ) that for any positive numbers @xmath6 and @xmath40 with @xmath41 , then @xmath128_{{\varepsilon},r}\mapsto { \operatorname{rank}}\kappa_0(p)-l\ ] ] is an isomorphism and @xmath129 . we have for any filtered @xmath2-algebra @xmath4 and any positive numbers @xmath6 , @xmath130 , @xmath40 and @xmath131 with @xmath132 and @xmath9 natural group homomorphisms called the structure maps : * @xmath133_{{\varepsilon},r}\mapsto [ \kappa_0(p)]-[i_l]$ ] ( where @xmath134 is the spectral projection associated to @xmath46 ) ; * @xmath135_{{\varepsilon},r}\mapsto [ u]$ ] ; * @xmath136 ; * @xmath137_{{\varepsilon},r}\mapsto [ p , l]_{{\varepsilon}',r'};$ ] * @xmath138_{{\varepsilon},r}\mapsto [ u]_{{\varepsilon}',r'}$ ] . * @xmath139 if some of the indices @xmath140 or @xmath141 are equal , we shall not repeat it in @xmath142 . the structures maps satisfy the obvious compatibilitity rules with respect to compositions . we have in the formalism of quantitative @xmath0-theory the analogue of the standard form for a @xmath0-theory class . [ lemma - almost - canonical - form ] let @xmath4 be a non unital filtered @xmath2-algebra . let @xmath40 and @xmath143 be positive numbers with @xmath144 . then for any @xmath18 in @xmath145 , there exist * two integers @xmath86 and @xmath5 with @xmath146 ; * @xmath53 a @xmath147-@xmath143-unitary in @xmath148 such that @xmath149 and @xmath150_{9 { \varepsilon } , s}$ ] in @xmath151 . let @xmath18 be an element in @xmath152 , let @xmath46 be an @xmath40-@xmath143-projection in some @xmath153 and let @xmath86 be an integer with @xmath154 and such that @xmath155_{{\varepsilon},s}$ ] . we can assume without loss of generality that @xmath156 . let @xmath157 be a unitary in @xmath158 such that @xmath159 . since @xmath157 is homotopic to @xmath160 as a unitary of @xmath158 , we see that @xmath161 and @xmath46 are homotopic as @xmath40-@xmath143-projections in @xmath153 . set then @xmath162 since @xmath163 we get according to lemma [ lemma - almost - closed ] that @xmath53 and @xmath164 are homotopic @xmath147-@xmath143-projections in @xmath153 . since @xmath165 , we get the result . we have a similar result in the odd case . [ lemma - almost - canonical - form - odd ] let @xmath4 be a non unital @xmath2-algebra filtered by @xmath166 . let @xmath40 and @xmath143 be positive numbers with @xmath167 . 1 . for any @xmath18 in @xmath168 , there exists an @xmath169-@xmath143-unitary @xmath42 in @xmath153 , such that @xmath170 and @xmath171_{21{\varepsilon},s}$ ] in @xmath172 ; 2 . if @xmath42 and @xmath57 are two of @xmath40-@xmath6-unitaries in @xmath153 such that @xmath173 and @xmath115_{{\varepsilon},s}=[v]_{{\varepsilon},s}$ ] in @xmath125 , then there exists an integer @xmath86 and a homotopy @xmath174}$ ] of @xmath169-@xmath143-unitaries of @xmath175 between @xmath176 and @xmath177 such that @xmath178 for every @xmath179 in @xmath180 $ ] . let @xmath57 be an @xmath40-@xmath6-unitary in some @xmath153 such that @xmath181_{{\varepsilon},r}$ ] . according to ( * remark 1.4 ) , we have that @xmath182 . in particular , @xmath183 is a @xmath184-@xmath6unitary and @xmath183 is homotopic to @xmath160 as a @xmath184-@xmath143-unitary of @xmath158 , where @xmath185 is provided with the trivial filtration ( * ? ? ? * lemma 1.20 ) . hence , if we set @xmath186 , then @xmath42 is a @xmath169-@xmath143 unitary of @xmath153 such that @xmath170 and @xmath42 and @xmath57 are homotopic as @xmath169-@xmath143 unitaries of @xmath153 . hence we have the equality @xmath187_{21{\varepsilon},s}=[u]_{21{\varepsilon},s}.\ ] ] if @xmath188 is a homomorphism filtered @xmath2-algebras , then since @xmath189 preserves @xmath40-@xmath6-projections and @xmath40-@xmath6-unitaries , it obviously induces for any positive number @xmath6 and any @xmath190 a group homomorphism @xmath191 moreover quantitative @xmath0-theory is homotopy invariant with respect to homotopies that preserves propagation ( * ? ? ? * lemma 1.26 ) . there is also a quantitative version of morita equivalence ( * ? ? ? * proposition 1.28 ) . [ prop - morita ] if @xmath4 is a filtered algebra and @xmath192 is a separable hilbert space , then the homomorphism @xmath193 induces a ( @xmath194-graded ) group isomorphism ( the morita equivalence ) @xmath195 for any positive number @xmath6 and any @xmath190 . the following observation establishes a connection between quantitative @xmath0-theory and classical @xmath0-theory ( see ( * ? ? ? * remark 1.17 ) ) . [ proposition - approximation - k - th ] 1 . let @xmath4 be a filtered @xmath1-algebra . for any positive number @xmath40 with @xmath196 and any element @xmath19 of @xmath197 , there exists a positive number @xmath6 and an element @xmath18 of @xmath198 such that @xmath199 ; 2 . there exists a positive number @xmath200 such that for any @xmath2-algebra @xmath4 , any positive numbers @xmath40 and @xmath6 with @xmath201 and any element @xmath18 of @xmath198 for which @xmath202 in @xmath197 , then there exists a positive number @xmath130 with @xmath203 such that @xmath204 in @xmath205 . as a consequence , we get the following approximation property . [ cor - limit - quant - k - groups ] let @xmath200 be as in proposition [ proposition - approximation - k - th ] . then for any positive number @xmath40 with @xmath206 and for any filtered @xmath2-algebra @xmath4 , then @xmath207 in order to study the functorial properties of quantitative @xmath0-theory , we introduced in @xcite the concept of quantitative object . a quantitative object is a family @xmath208 of abelian groups , together with a family of group homomorphisms @xmath209 for @xmath210 and @xmath211 called the structure maps such that * @xmath212 for any @xmath213 and @xmath214 ; * @xmath215 for any @xmath216 and @xmath217 . our prominent example will be of course quantitative @xmath0-theory @xmath218 of a filtered @xmath2-algebra @xmath4 with structure maps @xmath219 for @xmath210 and @xmath220 . in this subsection , we recall from ( * ? ? ? * section 2 ) the relevant notion of morphisms in the framework of quantitative objects . let @xmath71 be a control pair and let @xmath208 and @xmath221 be quantitative objects . a @xmath71-controlled morphism @xmath222 is a family @xmath223 of groups homomorphisms @xmath224 such that for any positive numbers @xmath225 and @xmath130 with @xmath226 and @xmath227 , we have @xmath228 when it is not necessary to specify the control pair , we will just say that @xmath229 is a controlled morphism . if @xmath208 is a quantitative object , let us define the identity @xmath230-controlled morphism @xmath231 recall that if @xmath4 and @xmath232 are filtered @xmath2-algebras and if @xmath233 is a @xmath71-controlled morphism , then @xmath229 induces a morphism @xmath234 uniquely defined by @xmath235 . in some situation , as for instance control boundary maps of controlled mayer - vietoris pair ( see section [ subs - section - boundary ] ) , we deal with family @xmath236 of group morphism defined indeed only up to a certain order . let @xmath71 be a control pair , let @xmath237 and @xmath238 be quantitative objects and let @xmath6 be a positive number . a @xmath71-controlled morphism of order @xmath6 @xmath222 is a family @xmath239 of groups homomorphisms @xmath240 such that for any positive numbers @xmath241 and @xmath242 with @xmath243 and @xmath244 , we have @xmath245 @xmath71 and @xmath246 are two control pairs , define @xmath247 then @xmath248 is again a control pair . let @xmath249 , @xmath250 and @xmath251 be quantitative objects , let @xmath252 be a @xmath253-controlled morphism , and let @xmath254 be a @xmath255-controlled morphism . then @xmath256 is the @xmath257-controlled morphism defined by the family @xmath258notice that if let @xmath259 and @xmath260 are respectively a @xmath253-controlled morphism and @xmath255-controlled morphism of order @xmath6 , then equation ( [ equ - composition ] ) defines a @xmath257-controlled morphism @xmath256 of order @xmath6 . let @xmath261 be a control pair and let @xmath208 and @xmath262be quantitative objects and let @xmath263 ( resp . @xmath264 ) be a @xmath253-controlled morphism ( resp . a @xmath255-controlled morphism ) . then we write @xmath265 if * @xmath266 and @xmath267 . * for every @xmath40 in @xmath268 and @xmath214 , then @xmath269 let @xmath270 and @xmath271 be controlled morphisms and let @xmath71 be a control pair . then the diagram ( or the square ) @xmath272 is called @xmath273 * -commutative * ( or @xmath274 * -commutes * ) if @xmath275 . the definition of @xmath71-commutative diagram can be obviously extended to the setting of controlled morphism of order @xmath6 . recall from @xcite the definition of a controlled isomorphisms . let @xmath71 be a control pair , and let @xmath259 be a @xmath253-controlled morphism with @xmath266 . @xmath276 is called @xmath71-invertible or a @xmath71-isomorphism if there exists a controlled morphism @xmath277 such that @xmath278 and @xmath279 . the controlled morphism @xmath280 is called a @xmath71-inverse for @xmath280 . in particular , if @xmath4 and @xmath232 are filtered @xmath2-algebras and if @xmath281 is a @xmath71-isomorphism , then the induced morphism @xmath282 is an isomorphism and its inverse is induced by a controlled morphism ( indeed induced by any @xmath71-inverse for @xmath229 ) . in order to state in section [ sec - kunneth ] the quantitative knneth formula , we will need the more general of quantitative isomorphism . let @xmath283 and @xmath284 be quantitative objects . for a @xmath285-controlled morphism @xmath286 consider the following statements : @xmath287 : : for any element @xmath18 in @xmath288 such that @xmath289 in @xmath290 , then @xmath291 in @xmath292 ; @xmath293 : : for any element @xmath19 in @xmath294 , there exists an element @xmath18 in @xmath295 such that @xmath296 in @xmath297 . [ def - quantitative - morphism ] * let @xmath283 and @xmath284 be quantitative objects . then a quantitative isomorphism @xmath298 is an @xmath285-controlled morphism @xmath299 for some control pair @xmath300 that satisfies the following : there exists a positive number @xmath301 , with @xmath302 such that * * for any positive numbers @xmath40 and @xmath143 with @xmath303 there exists a positive number @xmath242 with @xmath304 such that @xmath305 is satisfied ; * * for any positive numbers @xmath40 and @xmath143 with @xmath306 , there exists a positive number @xmath242 with @xmath307 such that @xmath308 is satisfied . the positive number @xmath301 is called the * rescaling * of the quantitative isomorphism @xmath229 . * a uniform family of quantitative isomorphisms if a family @xmath309 where , @xmath310 is for any @xmath311 in @xmath312 an @xmath300-controlled morphism for a given control pair @xmath300 such that there exists a positive number @xmath301 , with @xmath302 for which the following holds * * for any positive numbers @xmath40 and @xmath143 with @xmath303 there exists a positive number @xmath242 with @xmath304 such that @xmath313 is satisfied for any @xmath311 in @xmath312 ; * * for any positive numbers @xmath40 and @xmath143 with @xmath306 , there exists a positive number @xmath242 with @xmath314 such that @xmath315 is satisfied for any @xmath311 in @xmath312 . in particular , if @xmath4 and @xmath232 are filtered @xmath2-algebras and if @xmath281 is a quantitative isomorphism , then the induced morphism @xmath282 is an isomorphism ( but its inverse is no more given by a quantitative isomorphism ) . in this subsection , we recall the controlled exact sequence for quantitative objects . this controlled exact sequence is an important tool in computing quantitative @xmath0-theory for filtered @xmath1-algebras . let @xmath71 be a control pair , * let @xmath316 and @xmath317 be quantitative objects and let @xmath318 be a @xmath253-controlled morphism and let @xmath319 be a @xmath255-controlled morphism . then the composition @xmath320 is said to be @xmath71-exact at @xmath321 if @xmath322 and if for any @xmath323 , any @xmath324 and any @xmath19 in @xmath325 such that @xmath326 in @xmath327 , there exists an element @xmath18 in @xmath328 such that @xmath329 in @xmath330 . * a sequence of controlled morphisms @xmath331 is called @xmath71-exact if for every @xmath86 , the composition @xmath332 is @xmath71-exact at @xmath333 . * the notion of @xmath71-exactness of a composition and of a sequence can obviously be extended to the setting of controlled morphism of order @xmath6 . examples of controlled exact sequences in quantitative @xmath0-theory are provided by controlled six term exact sequences associated to a completely filtered extensions of @xmath2-algebras ( * ? ? ? * section 3 ) . [ def - completely - filtered - extension ] let @xmath4 be a @xmath2-algebra filtered by @xmath7 , let @xmath334 be an ideal of @xmath4 and set @xmath335 . the extension of @xmath2-algebras @xmath336 is called a completely filtered extension of @xmath2-algebras if the bijective continuous linear map @xmath337 induced by the inclusion @xmath338 is a complete isometry i.e for any integer @xmath5 , any positive number @xmath6 and any @xmath18 in @xmath339 , then @xmath340 notice that in this case , the ideal @xmath334 is filtered by @xmath341 and @xmath342 is filtered by @xmath343 . a particular case of completely filtered extension of @xmath2-algebra is the case of filtered and semi - split extension of @xmath2-algebras ( * ? ? ? * lemma 3.3 ) ( or a semi - split extension of filtered algebras ) i.e extension @xmath344 where * @xmath4 is filtered by @xmath7 ; * there exists a completely positive ( complete ) norm decreasing cross - section @xmath345 such that @xmath346 for any number @xmath214 . for any extension of @xmath2-algebras @xmath336 we denote by @xmath347 the associated ( odd degree ) boundary map in @xmath0-theory . [ prop - bound ] there exists a control pair @xmath348 such that for any completely filtered extension of @xmath2-algebras @xmath349 there exists a @xmath348-controlled morphism of odd degree @xmath350 which induces in @xmath0-theory @xmath351 . moreover the controlled boundary map enjoys the usual naturally properties with respect to extensions ( see ( * ? ? ? * remark 3.8 ) ) . [ thm - six term ] there exists a control pair @xmath71 such that for any completely filtered extension of @xmath2-algebras @xmath352 then the following six - term sequence is @xmath71-exact @xmath353 in this subsection , we discuss compatibility of kasparov s @xmath354-theory with quantitative @xmath0-theory of filtered @xmath1-algebras . let @xmath4 be a @xmath2-algebra and let @xmath232 be a filtered @xmath2-algebra filtered by @xmath355 . let us define @xmath356 as the closure in the spatial tensor product @xmath357 of the algebraic tensor product of @xmath4 and @xmath358 . then the @xmath2-algebra @xmath357 is filtered by @xmath359 . if @xmath360 is a homomorphism of @xmath2-algebras , let us set @xmath361 recall from @xcite that for @xmath2-algebras @xmath362 , @xmath363 and @xmath232 , kasparov defined a tensorization map @xmath364 if @xmath232 is a filtered @xmath2-algebra , then for any @xmath365 in @xmath366 the morphism @xmath367 is induced by a controlled morphism which enjoys compatibity properties with kasparov product ( * ? ? ? * theorem 4.4 ) . [ thm - tensor ] there exists a control pair @xmath368 such that * for any filtered @xmath2-algebra @xmath232 ; * for any @xmath2-algebras @xmath362 and @xmath363 ; * for any element @xmath365 in @xmath366 , there exists a @xmath368-controlled morphism @xmath369 with same degree as @xmath365 that satisfies the following : 1 . @xmath369 induces in @xmath0-theory the right multiplication by @xmath370 ; 2 . for any elements @xmath365 and @xmath371 in @xmath366 then @xmath372 3 . let @xmath373 be a filtered @xmath2-algebras and let @xmath374 be a homomorphism of @xmath2-algebras , then @xmath375 for all @xmath365 in @xmath376 . 4 . let @xmath377 be a @xmath2-algebra and let @xmath378 be a homomorphism of @xmath2-algebras then @xmath379 for any @xmath365 in @xmath380 . @xmath381)\stackrel{(\alpha_{\mathcal{t}},k_{\mathcal{t}})}{\sim } { \mathcal{i}d}_{{\mathcal{k}}_*({a_1}{{\otimes}}b)}$ ] . 6 . for any @xmath2-algebra @xmath382 and any element @xmath365 in @xmath366 , we have @xmath383 . 7 . for any semi - split extension of @xmath2-algebras @xmath384 with corresponding element @xmath385 $ ] of @xmath386 that implements the boundary map , then we have @xmath387)\stackrel{({{\alpha}}_{\mathcal{t}},k_{\mathcal{t}})}{\sim}{\mathcal d}_{j{{\otimes}}b , a{{\otimes}}b}.\ ] ] moreover , the controlled tensorization morphism @xmath388 is compatible with kasparov products . [ thm - product - tensor ] there exists a control pair @xmath71 such that the following holds : let @xmath389 and @xmath390 be separable @xmath2-algebras and let @xmath232 be a filtered @xmath2-algebra . then for any @xmath365 in @xmath366 and any @xmath371 in @xmath391 , we have @xmath392 we also have in the case of finitely generated group a controlled version of the kasparov transformation . let @xmath393 be a finitely generated group . recall that a length on @xmath394 is a map @xmath395 such that * @xmath396 if and only if @xmath397 is the identity element @xmath398 of @xmath394 ; * @xmath399 for all element @xmath397 and @xmath400 of @xmath394 . * @xmath401 . in what follows , we will assume that @xmath402 is a word length arising from a finite generating symmetric set @xmath403 , i.e. , @xmath404 . let us denote by @xmath405 the ball centered at the neutral element of @xmath393 with radius @xmath6 , i.e. , @xmath406 . let @xmath4 be a separable @xmath393-@xmath2-algebra , i.e. , a separable @xmath2-algebra provided with an action of @xmath393 by automorphisms . for any positive number @xmath6 , we set @xmath407 then the @xmath2-algebra @xmath408 is filtered by @xmath409 . moreover if @xmath410 is a @xmath393-equivariant morphism of @xmath2-algebras , then the induced homomorphism @xmath411 is a filtered homomorphism . in @xcite , for any @xmath393-@xmath2-algebras @xmath4 and @xmath232 , was constructed a natural transformation @xmath412 that preserves kasparov products , called the kasparov transformation . the kasparov transformation admits a quantitative version ( * ? ? ? * section 5 ) . [ thm - kas ] there exists a control pair @xmath413 such that * for any separable @xmath394-@xmath2-algebras @xmath4 and @xmath232 ; * for any @xmath365 in @xmath414 , there exists a @xmath413-controlled morphism @xmath415 of same degree as @xmath365 that satisfies the following : 1 . for any element @xmath365 of @xmath414 , then @xmath416 induces in @xmath0-theory right multiplication by @xmath417 . 2 . for any @xmath365 and @xmath371 in @xmath418 , then @xmath419 3 . for any @xmath394-@xmath2-algebra @xmath28 , any homomorphism @xmath420 of @xmath394-@xmath2-algebras and any @xmath365 in @xmath421 , then @xmath422 . 4 . for any @xmath394-@xmath2-algebra @xmath423 , any homomorphism @xmath424 of @xmath394-@xmath2-algebras and any @xmath365 in @xmath418 , then @xmath425 . 5 . @xmath426 . if @xmath336 is a semi - split exact sequence of @xmath393-@xmath2-algebras , let @xmath385 $ ] be the element of @xmath427 that implements the boundary map @xmath428 . then we have @xmath429)\stackrel{({{\alpha}}_{\mathcal{j}},k_{\mathcal{j}})}{\sim}{\mathcal d}_{j{{\rtimes_{red}}}{{\gamma}},a{{\rtimes_{red}}}{{\gamma}}}.\ ] ] the controlled kasparov transformation is compatible with kasparov products . [ thm - product ] there exists a control pair @xmath71 such that the following holds : for every separable @xmath393-@xmath2-algebras @xmath430 and @xmath382 , any elements @xmath365 in @xmath418 and @xmath371 in @xmath431 , then @xmath432 [ remark - kasp - tens ] we can choose @xmath433 such that @xmath434 . in this case , for any @xmath393-@xmath2-algebra @xmath4 , any @xmath2-algebras @xmath435 and @xmath436 equipped with the trivial action of @xmath393 and any @xmath365 in @xmath437 , then @xmath438 we have similar result for maximal crossed products . in this subsection , we discuss a quantitative version of the baum - connes assembly map . let @xmath393 be a finitely generated group and let @xmath232 be a @xmath393-@xmath2-algebra . we equip @xmath393 with any word metric . recall that if @xmath439 is a positive number , then the rips complex of degree @xmath439 is the set @xmath440 of probability measures with support of diameter less than @xmath439 . then @xmath440 is a locally finite simplicial complex and provided with the simplicial topology , @xmath440 is endowed with a proper and cocompact action of @xmath393 by left translation . in @xcite , for any @xmath393-@xmath2-algebra @xmath232 , we construct quantitative assembly maps @xmath441 with @xmath442 and @xmath443 , where @xmath444 is a function independent on @xmath232 , non decreasing in @xmath439 and non increasing in @xmath40 . moreover , the maps @xmath445 induce the usual assembly maps @xmath446 i.e. , @xmath447 . let us recall now the definition of the quantitative assembly maps . observe first that any @xmath18 in @xmath440 can be written down in a unique way as a finite convex combination @xmath448 where @xmath449 is the dirac probability measure at @xmath450 in @xmath393 . the functions @xmath451\ ] ] are continuous and @xmath452 for all @xmath450 and @xmath453 in @xmath393 . the function @xmath454 is a projection of @xmath455 with propagation less than @xmath439 . let us set then @xmath456 , where the control pair @xmath433 is as intheorem [ thm - kas ] . recall that @xmath457 can be chosen non increasing and in this case , @xmath458 is non decreasing in @xmath439 and non increasing in @xmath40 . [ def - quantitative - assembly - map ] for any @xmath394-@xmath2-algebra @xmath4 and any positive numbers @xmath40 , @xmath6 and @xmath439 with @xmath41 and @xmath443 , we define the quantitative assembly map @xmath459_{\frac{{\varepsilon}}{\alpha_{\mathcal{j}}},\frac{r}{k_{{\mathcal{j}},{{\varepsilon}}/{\alpha_{\mathcal{j}}}}}}\right),\end{aligned}\ ] ] where the notation @xmath460_{\frac{{\varepsilon}}{\alpha_{\mathcal{j}}},\frac{r}{k_{{\mathcal{j}},{{\varepsilon}}/{\alpha_{\mathcal{j}}}}}}$ ] is as in definition 1.5 . then according to theorem [ thm - kas ] , the map @xmath461 is a group homomorphism . for any positive numbers @xmath439 and @xmath462 such that @xmath463 , we denote by @xmath464 the homomorphism induced by the restriction from @xmath465 to @xmath440 . it is straightforward to check that if @xmath439 , @xmath462 and @xmath6 are positive numbers such that @xmath463 and @xmath466 , then @xmath467 . moreover , for every positive numbers @xmath468 and @xmath130 such that @xmath469 , @xmath470 , @xmath471 , and @xmath472 , we get by definition of a controlled morphism that @xmath473 in @xcite we introduced quantitative statements for the quantitative assembly maps . for a @xmath393-@xmath2-algebra @xmath4 and positive numbers @xmath474 and @xmath131 with @xmath463 , @xmath475 , @xmath476 and @xmath477 we set : @xmath478 : : for any element @xmath18 in @xmath479 , then @xmath480 in @xmath481 implies that @xmath482 in @xmath483 . @xmath484 : : for every @xmath19 in @xmath481 , there exists an element @xmath18 in @xmath479 such that @xmath485 the following results were then proved . [ thm - quant - surj]let @xmath393 be a discrete group . 1 . assume that for any @xmath393-@xmath2-algebra @xmath4 , the assembly map @xmath486 is one - to - one . then for any positive numbers @xmath439 , @xmath40 and @xmath443 with @xmath41 and @xmath487 , there exists a positive number @xmath462 with @xmath488 such that @xmath489 is satisfied for every @xmath393-@xmath2-algebra @xmath4 ; 2 . assume that for any @xmath393-@xmath2-algebra @xmath4 , the assembly map @xmath486 is onto . then for some positive number @xmath490 which not depends on @xmath393 or on @xmath4 and such that with @xmath491 and for any positive numbers @xmath40 and @xmath6 with @xmath492 , there exist positive numbers @xmath439 and @xmath130 with @xmath493 and @xmath9 such that @xmath494 is satisfied for every @xmath393-@xmath2-algebra @xmath4 . in particular , if @xmath393 satisfies the baum - connes conjecture with coefficients , then @xmath393 satisfies points ( i ) and ( ii ) above . in @xcite we developed a geometric version of the controlled assembly maps and of the quantitative statements in the following setting . let @xmath495 be a proper discrete metric space and let @xmath4 be a @xmath2-algebra . then the distance @xmath439 on @xmath495 induces a filtration on @xmath496 in the following way : let @xmath6 be a positive number and @xmath497 be an element in @xmath496 , with @xmath498 in @xmath4 for any @xmath499 and @xmath500 in @xmath501 . then @xmath502 has propagation less that @xmath6 if @xmath503 for @xmath499 and @xmath500 in @xmath495 such that @xmath504 . as for finitely generated group , we define the rips complex of degree @xmath439 of @xmath495 as the set @xmath505 of probability measure with support of diameter less than @xmath439 . then @xmath505 is a locally finite simplicial complex and is locally compact when endowed with the simplicial topology . let us define then @xmath506 where @xmath507 runs through compact subset of @xmath505 . in @xcite , for any @xmath2-algebra @xmath4 , was constructed local quantitative coarse assembly maps @xmath508 with @xmath442 and @xmath443 . the map @xmath509 is natural in the @xmath2-algebra and induces in @xmath0-theory the index map , i.e the maps @xmath510 is up to morita equivalence given for any compact subset @xmath507 of @xmath505 by the morphism in the inductive limit @xmath511 induced by the map @xmath512 . moreover , the maps @xmath513 are compatible with structure morphisms and with inclusion of rips complexes : * @xmath514 for any positive numbers @xmath515 and @xmath143 such that @xmath516 and @xmath9 ; * @xmath517 for any positive numbers @xmath518 and @xmath462 such that @xmath519 and @xmath520 , where @xmath521 is the homomorphism induced by the restriction from @xmath522 to @xmath505 let us consider then for @xmath474 and @xmath131 positive numbers with @xmath463 , @xmath523 , @xmath470 and @xmath524 the following statements : @xmath525 : : for any element @xmath18 in @xmath526 , then @xmath527 in @xmath528 implies that @xmath482 in @xmath529 . @xmath530 : : for every @xmath19 in @xmath531 , there exists an element @xmath18 in @xmath526 such that @xmath532 recall that a a proper discrete metric space @xmath495 with bounded geometry coarsely embeds in a hilbert space if there exist * a map @xmath533 where @xmath192 is a hilbert space ; * @xmath534 two maps with @xmath535 , such that @xmath536 for any @xmath18 and @xmath19 in @xmath495 . proper discrete metric spaces with bounded geometry that coarsely embed into a hilbert space provide numerous examples that satisfy the above geometric quantitative statement ( * ? ? ? * theorems 4.9 and 4.10 ) . [ thm - geo - statement ] let @xmath495 be a discrete metric space with bounded geometry that coarsely embeds into a hilbert space . 1 . for any positive numbers @xmath439 , @xmath40 and @xmath6 with @xmath41 and @xmath443 , there exists a positive number @xmath462 with @xmath488 for which @xmath525 is satisfied for any @xmath2-algebra @xmath4 . 2 . there exists a positive number @xmath537 such that for any positive numbers @xmath40 and @xmath130 with @xmath538 , there exist positive numbers @xmath439 and @xmath6 with @xmath539 and @xmath540 for which @xmath541 is satisfied for any @xmath2-algebra @xmath4 . in this section , we introduce the concepts of neighborhood @xmath1-algebras and controlled mayer - vietoris pair . this is the cornerstone for the controlled mayer - vietoris sequence for quantitative @xmath0-theory which will be stated in section [ section - six - terms ] . we also discuss a few technical lemmas useful for establishing the latter . in ( * ? ? ? * section @xmath543 ) was introduced the notion of @xmath40-@xmath6-@xmath542-invertible element of a unital banach algebra . in this subsection , we study @xmath40-@xmath6-@xmath542-invertible elements for @xmath2-algebras . in particular , we state an analogue of the polar decomposition in the setting of @xmath40-@xmath6-unitaries . let @xmath4 be a unital @xmath2-algebra filtered by @xmath166 and let @xmath544 and @xmath542 be positive numbers with @xmath545 . an element @xmath18 in @xmath10 is called @xmath40-@xmath6-@xmath542-invertible if @xmath546 and there exists @xmath19 in @xmath10 such that @xmath547 and @xmath548 . such an element @xmath19 is called an @xmath40-@xmath6-@xmath542-inverse for @xmath18 . [ rem - norm - inv]if @xmath18 is @xmath40-@xmath6-@xmath542-invertible , then @xmath18 is invertible and for any @xmath40-@xmath6-@xmath542-inverse @xmath19 for @xmath18 , we have @xmath549 . let @xmath4 be a unital @xmath2-algebra filtered by @xmath166 and let @xmath544 and @xmath542 be positive numbers with @xmath545 . two @xmath40-@xmath6-@xmath542-invertibles in @xmath4 are called homotopic if there exists @xmath550\to a$ ] an @xmath40-@xmath6-@xmath542-invertible in @xmath551 $ ] such that @xmath552 and @xmath553 . in the setting of @xmath40-@xmath6-@xmath542-invertibles and of @xmath40-@xmath6-unitaries , there is the analogue of the polar decomposition . [ lemma - polar - decomp ] for any positive number @xmath542 there exists a control pair @xmath554 and a positive number @xmath555 with @xmath556 such that the following holds . for any filtered unital @xmath2-algebra @xmath4 filtered by @xmath166 , any positive numbers @xmath40 and @xmath6 with @xmath306 and every @xmath40-@xmath6-@xmath542-invertible element @xmath18 of @xmath4 , there exist @xmath104 positive and @xmath88-@xmath557-@xmath555-invertible in @xmath4 and @xmath42 an @xmath88-@xmath557-unitary in @xmath4 such that @xmath558 and @xmath559 . moreover we can choose @xmath42 and @xmath104 such that * there exists a real polynomial function @xmath560 with @xmath561 such that @xmath562 and @xmath563 ; * @xmath104 admits a positive @xmath88-@xmath557-@xmath555-inverse ; * if @xmath18 is homotopic to @xmath13 as an @xmath40-@xmath6-@xmath542-invertible , then @xmath42 is homotopic to @xmath13 as an @xmath88-@xmath557-unitary . according to remark [ rem - norm - inv ] and since @xmath41 , if @xmath18 is an @xmath40-@xmath6-@xmath542-invertible , then @xmath18 is invertible and @xmath564 and hence @xmath565 . this implies that the spectrum of @xmath566 is included in @xmath567 $ ] . let @xmath568 and @xmath569 be positive numbers such that @xmath570 and @xmath571 . let us consider the power serie @xmath572 of @xmath573 for @xmath179 in @xmath180 $ ] and let @xmath574 be the smallest integer such that @xmath575 and @xmath576 for all @xmath179 in @xmath577 $ ] . since @xmath578 converges to @xmath579 , if we set @xmath580 then @xmath560 is a polynomial of degree @xmath574 such that @xmath581 @xmath582 for every @xmath179 in @xmath577 $ ] and @xmath583 . if we set @xmath562 , then @xmath42 is a @xmath584-@xmath585-unitary for some @xmath586 depending only on @xmath542 . set now @xmath587 , then up to taking a large @xmath584 , there exists a control pair @xmath81 and a positive number @xmath555 depending only on @xmath542 , with @xmath556 and such that @xmath42 and @xmath104 satisfy the required properties and @xmath588 is a positive @xmath88-@xmath589-@xmath555-inverse for @xmath104 . moreover , if @xmath590}$ ] is a homotopy of @xmath40-@xmath6-invertibles between @xmath13 and @xmath18 , then @xmath591}$ ] is a homotopy of @xmath88-@xmath589-unitaries between @xmath13 and @xmath42 . let @xmath4 be a unital @xmath2-algebra filtered by @xmath166 . for @xmath18 and @xmath19 in @xmath4 , set @xmath592 and @xmath593 and consider the commutators @xmath594 and @xmath595 [ lem - commutator ] let @xmath4 be a unital @xmath2-algebra filtered by @xmath166 and let @xmath40 and @xmath6 be positive numbers with @xmath41 . let @xmath596 and @xmath597 in @xmath10 such that @xmath598 is an @[email protected] we have the inequality @xmath599 let us set @xmath600 . consider the matrix @xmath601 since @xmath42 is an @xmath40-@xmath6-unitary , then @xmath602 we have @xmath603 this , together with the definition of @xmath507 and @xmath604 , implies that @xmath605 we introduce in this subsection the basic ingredient to define controlled mayer - vietoris pairs . if @xmath607 and @xmath608 are two closed linear subspace of a @xmath2-algebra @xmath4 such that @xmath609 , we equipped @xmath610 with the quotient @xmath2-algebra norm , i.e if @xmath18 is a element of @xmath611 , then @xmath612 . then this family of norms is a matrix norm on @xmath613 . let @xmath4 be a @xmath2-algebra filtered by @xmath166 and let @xmath6 be a positive number . * a coercive decomposition pair of degree @xmath6 for @xmath4 ( or a coercive decomposition @xmath6-pair ) is a pair @xmath614 of closed linear subspaces of @xmath10 such that there exists a positive number @xmath615 satisfying the following : for any positive number @xmath143 with @xmath616 the inclusion @xmath617 induces an isomorphism @xmath618 whose inverse is bounded in norm by @xmath615 . * a completely coercive decomposition pair of degree @xmath6 for @xmath4 ( or a completely coercive decomposition @xmath6-pair ) is a pair @xmath614 of closed linear subspaces of @xmath10 such that there exists a positive number @xmath615 satisfying the following : for any positive number @xmath143 with @xmath616 the inclusion @xmath617 induces a complete isomorphism @xmath618 whose inverse has complete norm bounded by @xmath615 . let @xmath4 be a @xmath2-algebra filtered by @xmath166 , let @xmath6 be a positive number and let @xmath614 be a pair of closed linear subspaces of @xmath10 . then @xmath614 is a coercive decomposition pair of degree @xmath6 for @xmath4 if and only if there exists a positive number @xmath619 such that for every positive number @xmath143 with @xmath616 and any @xmath18 in @xmath620 , there exists @xmath596 in @xmath621 and @xmath597 in @xmath622 , both with norm at most @xmath623 and such that @xmath624 . in the same way , @xmath614 is a completely coercive decomposition pair of degree @xmath6 for @xmath4 if and only if there exists a positive number @xmath619 such that for every positive number @xmath143 with @xmath616 , any integer @xmath5 and any @xmath18 in @xmath625 , there exists @xmath596 in @xmath626 and @xmath597 in @xmath627 , both with norm at most @xmath623 and such that @xmath624 . the ( completely ) coercive decomposition @xmath6-pair @xmath628 is said to have * coercity @xmath629*. the aim of this subsection is to show that for any coercive decomposition @xmath6-pair @xmath628 , there exists a control pair @xmath300 depending indeed only on the coercitivity , such that up to stabilisation , any @xmath40-@xmath143-unitary of @xmath4 with @xmath630 and @xmath631 can be decomposed as a product of two @xmath632-@xmath633-unitaries lying respectively in the unitarization of some suitable neighborhood algebras of @xmath634 and @xmath635 . let @xmath4 be a @xmath2-algebra filtered by @xmath166 . let @xmath6 and @xmath606 be positive numbers and let @xmath636 be a closed linear subspace of @xmath10 . we define @xmath637 , the @xmath606-neighborhood @xmath2-algebra of @xmath636 , as the @xmath1-subalgebra of @xmath4 generated by its @xmath606-neighborhood @xmath638 . notice that @xmath637 inherits from @xmath4 a structure of filtered @xmath2-algebra with @xmath639 for every positive number @xmath143 . for a positive number @xmath143 satisfying @xmath616 , we also denote by @xmath640 for the @xmath606-neighbohood @xmath2-algebra of @xmath641 . [ lemma - decomp - invertible ] for any positive number @xmath619 , there exist positive numbers @xmath642 and @xmath542 , with @xmath537 and @xmath643 such that the following holds . let @xmath4 be a unital @xmath2-algebra filtered by @xmath166 , let @xmath6 and @xmath40 be positive numbers such that @xmath538 and let @xmath614 be a coercive decomposition pair for @xmath4 of degree @xmath6 with coercity @xmath619 . then for any @xmath40-@xmath6-unitary @xmath42 in @xmath4 homotopic to @xmath13 , there exist an integer @xmath86 and @xmath644 and @xmath645 in @xmath646 such that * @xmath644 and @xmath645 are invertible ; * @xmath647 and @xmath648 are elements in the matrix algebra @xmath649 for @xmath650 ; * @xmath651 and @xmath652 for @xmath650 ; * for @xmath650 , there exists a homotopy @xmath653}$ ] of invertible elements in @xmath646 between @xmath13 and @xmath654 such that @xmath655 and @xmath656 and @xmath657 are elements in the matrix algebra @xmath649 for every @xmath179 in @xmath180 $ ] . * @xmath658 . let @xmath659}$ ] be a homotopy of @xmath40-@xmath6-unitaries of @xmath4 between @xmath660 and @xmath661 and let @xmath662 be a partition of @xmath180 $ ] such that @xmath663 for @xmath664 . set @xmath665 and @xmath666 then we have @xmath667 for any matrix @xmath668 in @xmath669 , let us set @xmath670 in @xmath671 . for every integer @xmath672 , pick @xmath673 in @xmath634 and @xmath674 in @xmath635 such that @xmath675 such that @xmath676 and @xmath677 . set @xmath678 and @xmath679 . since we have @xmath680 and @xmath681 then if we set @xmath682 we deduce from lemma [ lem - commutator ] that @xmath683 with @xmath684 in @xmath685 and hence @xmath686 let us show that @xmath687 can be decomposed as a product @xmath688 with @xmath644 and @xmath645 satisfying the required properties . notice that as a product of elementary matrices , then @xmath689 is invertible . by the definition of the neighborhood @xmath1-algebra , @xmath690 and @xmath691 are elements in the matrix algebra @xmath692 . the same holds for @xmath693 , and we have similar properties for @xmath694 and @xmath695 with respect to @xmath696 . in order to bring out some commutators , let us study the term @xmath697 which is the product of the four first terms in equation ( [ equation - s ] ) . let us make the following observations : * @xmath698 is an element in the matrix algebra @xmath699 ; * @xmath700 is an element in the matrix algebra @xmath701 ( because @xmath702 and @xmath703 are elements of the matrix algebra @xmath704 ) ) ; * @xmath705 is an element in the matrix algebra @xmath706 ; * @xmath707 is an element in the matrix algebra @xmath708 . hence @xmath709 is an element in the matrix algebra @xmath710 . since @xmath711 , then we have @xmath712 therefore there exists for @xmath650 an invertible matrix @xmath713 in @xmath671 such that @xmath714 and @xmath715 lie in @xmath716and @xmath717 therefore if we write @xmath718 with @xmath719 and @xmath720 are invertible matrices of @xmath671 such that * @xmath721 and @xmath722 are elements in the matrix algebra @xmath723 . * @xmath724 and @xmath725 are elements in the matrix algebra @xmath726 . since @xmath644 and @xmath645 can be written as a product of a fixed number , say @xmath46 , of matrices @xmath727 or @xmath728 with @xmath729 , we see that @xmath644 and @xmath645 have norm less than @xmath730 . according to equation ( [ equation - p1p2 ] ) , we have @xmath731 the required homotopies are then @xmath732}\ ] ] and @xmath733}.\ ] ] let us briefly explain how we deal with the non unital case . let @xmath4 be a non unital filtered @xmath2-algebra and let @xmath42 be an @xmath40-@xmath6 unitary in @xmath32 such that @xmath734 is in @xmath4 . assume that @xmath600 with @xmath735 and @xmath597 lie in @xmath4 . proceeding as in the proof of lemma [ lem - commutator ] , we see that @xmath69 is @xmath66-close to the product of @xmath736 and @xmath737 now if @xmath614 be a coercive decomposition pair for @xmath4 of degree @xmath6 with coercity @xmath619 and assume that in the above decomposition of @xmath600 we have @xmath735 in @xmath634 and @xmath597 in @xmath635 , we get then by a straightforward computation that @xmath738 has coefficient in @xmath739 and @xmath740 has coefficient in @xmath741 . notice that in view of the proof of lemma [ lemma - almost - canonical - form - odd ] , if under above assumption , @xmath42 is connected to @xmath13 as a @xmath40-@xmath6-unitary of @xmath32 , then @xmath42 is connected to @xmath13 by a homotopy of @xmath169-@xmath6-unitaries @xmath659}$ ] of @xmath32 such that @xmath742 lies in @xmath4 for all @xmath179 in @xmath180 $ ] . hence proceeding as in the proof of lemma [ lemma - decomp - invertible ] we get : [ lemma - decomp - invertible - non - unital ] for any positive number @xmath619 , there exist positive numbers @xmath642 and @xmath542 , with @xmath537 and @xmath643 such that the following holds . let @xmath4 be a non unital @xmath2-algebra filtered by @xmath166 , let @xmath6 and @xmath40 be positive numbers such that @xmath538 and let @xmath614 be a coercive decomposition pair for @xmath4 of degree @xmath6 with coercity @xmath619 . then for any @xmath40-@xmath6-unitary @xmath42 in @xmath32 homotopic to @xmath13 and such that @xmath734 lies in @xmath4 , there exist an integer @xmath86 and @xmath644 and @xmath645 in @xmath743 such that * @xmath644 and @xmath645 are invertible ; * @xmath647 and @xmath648 are elements in the matrix algebra @xmath744 for @xmath650 ; * @xmath651 and @xmath652 for @xmath650 ; * for @xmath650 , there exists a homotopy @xmath653}$ ] of invertible elements in @xmath743 between @xmath13 and @xmath654 such that @xmath655 and @xmath656 and @xmath657 are elements in the matrix algebra @xmath744 for every @xmath179 in @xmath180 $ ] . * @xmath658 . [ proposition - decom - unitaries ] for every positive number @xmath619 , there exists a control pair @xmath554 such that the following holds . let @xmath4 be a unital @xmath2-algebra filtered by @xmath166 , let @xmath6 and @xmath40 be positive numbers such that @xmath82 and let @xmath614 be a coercive decomposition pair for @xmath4 of degree @xmath6 with coercity @xmath619 . then for any @xmath40-@xmath6-unitary @xmath42 in @xmath4 homotopic to @xmath13 , there exist a positive integer @xmath86 and @xmath745 and @xmath746 two @xmath88-@xmath557-unitaries in @xmath747 such that * @xmath748 is an element in the matrix algebra @xmath749 for @xmath650 ; * for @xmath650 , there exists a homotopy @xmath750}$ ] of @xmath88-@xmath557-unitaries between @xmath13 and @xmath674 such that @xmath751 for all @xmath179 in @xmath180 $ ] . * @xmath752 . as in lemma [ lemma - decomp - invertible ] , let @xmath642 and @xmath542 be positive numbers , @xmath86 be an integer and @xmath644 and @xmath645 be matrices of @xmath646 such that @xmath658 . since @xmath644 and @xmath645 are @xmath40-@xmath753-@xmath542-invertible for every @xmath40 in @xmath754 , then according to lemma [ lemma - polar - decomp ] , there exists * a control pair @xmath554 ; * @xmath745 an @xmath40-@xmath755-unitary and @xmath756 an @xmath40-@xmath755-@xmath542-invertible both in @xmath757 with @xmath756 positive and admitting a positive @xmath40-@xmath755-@xmath542-inverse ; * @xmath746 an @xmath40-@xmath755-unitary and @xmath758 an @xmath40-@xmath759-@xmath542-invertible both in @xmath760 with @xmath758 positive and admitting a positive @xmath40-@xmath755-@xmath542-inverse , such that @xmath761 and @xmath762 . then @xmath763 and hence , up to replacing @xmath72 by @xmath764 , we know according to lemma [ lemma - almost - closed ] that @xmath765 is a @xmath766-@xmath767-unitary . let us prove that @xmath768 is closed to @xmath769 . let @xmath770 be a positive @xmath40-@xmath755-@xmath542-inverse for @xmath756 . then we have @xmath771 and then @xmath772 . this implies that @xmath773 since @xmath774 and @xmath775 , we deduce that there exists @xmath776 depending only on @xmath72 and @xmath542 such that @xmath777 . but , since @xmath778 and @xmath758 are @xmath40-@xmath755-@xmath542-invertible with @xmath779 , their spectrum is bounded below by @xmath780 . the square root is lipschitz on the set of positive elements of @xmath4 with spectrum bounded below by @xmath780 ( this can be checked easily by holomorphic functional calculus ) , thus there exists a positive number @xmath781 , depending only on @xmath542 such that @xmath782 . since @xmath778 is an @xmath40-@xmath755-@xmath542-inverse for @xmath756 , we finally obtain that @xmath783 . combining this inequality with equation ( [ equ - lemma - polar ] ) , we know that there exist a positive number @xmath784 , depending only on @xmath542 and @xmath785 such that @xmath786 according to lemma [ lemma - polar - decomp ] , @xmath787 where @xmath560 is polynomial and such that @xmath561 . since @xmath721 lies in @xmath788 , then the same holds for @xmath789 and similarly , @xmath790 lies in @xmath791 . proceeding similarly , lemma [ lemma - decomp - invertible - non - unital ] allows to deal with the non unital case . [ proposition - decom - unitaries - non - unital ] for every positive number @xmath619 , there exists a control pair @xmath554 such that the following holds . let @xmath4 be a non unital @xmath2-algebra filtered by @xmath7 , let @xmath6 and @xmath40 be positive numbers such that @xmath82 and let @xmath614 be a coercive decomposition pair for @xmath4 of degree @xmath6 with coercity @xmath619 . then for any @xmath40-@xmath6-unitary @xmath42 in @xmath32 homotopic to @xmath13 and such that @xmath734 lies in @xmath4 , there exist a positive integer @xmath86 and @xmath745 and @xmath746 two @xmath88-@xmath557-unitaries in @xmath792 such that * @xmath748 is an element in the matrix algebra @xmath793 for @xmath650 ; * for @xmath650 , there exists a homotopy @xmath750}$ ] of @xmath88-@xmath557-unitaries between @xmath13 and @xmath674 such that @xmath794 for all @xmath179 in @xmath180 $ ] . * @xmath752 . in this subsection , we give the definition of neighborhood @xmath1-algebras and controlled mayer - vietoris pair . our prominent examples will be given by roe algebras . let @xmath4 be a @xmath2-algebra filtered by @xmath166 , let @xmath6 be a positive number and let @xmath636 be a closed linear subspace of @xmath10 . then a sub-@xmath2-algebra @xmath232 of @xmath4 is called an @xmath6-controlled @xmath636-neighborhood-@xmath2-algebra if * @xmath232 is filtered by @xmath795 ; * @xmath796 . [ example - roe - algebras ] let @xmath495 be a discrete metric space with bounded geometry and consider @xmath797 the roe algebra of @xmath495 . recall that @xmath797 is the closure of the algebra of locally compact and finite propagation operators on @xmath798 , where @xmath192 is a fixed separable hilbert space . then @xmath797 is filtered by the propagation . for @xmath6 a positive number , let @xmath799 be a family of finite subsets of @xmath495 whith uniformly bounded diameter which is @xmath606-disjoint ( i.e. , @xmath800 if @xmath801 ) for some positive number @xmath802 . let us consider the set @xmath803 of locally compact operators on @xmath798 with support in @xmath804 for a positive number @xmath143 , let us set @xmath805 . if @xmath806 , then @xmath807 is a family of @xmath808-disjoint subsets of @xmath495 with uniformly bounded diameter . consider then the subalgebra @xmath809 of @xmath797 of operators with support in @xmath810 . then @xmath811 and @xmath809 is for every @xmath143 with @xmath812 an @xmath6-controlled @xmath636-neighborhood-@xmath2-algebra . [ def - cia ] let @xmath813 and @xmath814 be two subsets of a @xmath2-algebra @xmath4 . the pair @xmath815 is said to have _ complete intersection approximation _ property ( cia ) if there exists @xmath816 such that for any positive number @xmath40 , any @xmath817 and @xmath818 for some @xmath5 and @xmath819 , then there exists @xmath820 satisfying @xmath821 the positive number @xmath619 is called the * coercity * of the pair @xmath815 . in the above definition , we note that the inequalities @xmath819 and @xmath822 implies @xmath823 hence we can remove the condition @xmath824 up to replacing the constant @xmath619 by @xmath825 . [ definition - weak - coarse - mv - pair ] let @xmath4 be a @xmath2-algebra filtered by @xmath166 and let @xmath6 be a positive number . an @xmath6-controlled weak mayer - vietoris pair for @xmath4 is a quadruple @xmath826 such that for some positive number @xmath619 . 1 . @xmath628 is a completely coercive decomposition pair for @xmath4 of order @xmath6 with coercitivity @xmath619 . @xmath827 is an @xmath6-controlled @xmath828-neighborhood-@xmath2-algebra for @xmath650 ; 3 . the pair @xmath829 has the cia property with coercitivity @xmath619 as defined above for any positive number @xmath143 with @xmath616 . the positive number @xmath619 is called the * coercity * of the @xmath6-controlled weak mayer - vietoris pair @xmath826 . in the above definition , 1 . @xmath830 is an @xmath143-controlled mayer - vietoris pair for any @xmath831 with same coercitivity as @xmath826 . @xmath832 is filtered by @xmath833 . in order to ensure some persistence properties for the controlled mayer - vietoris exact sequence ( see corollary [ cor - bound1 ] ) , we need to strenghen condition ( iii ) of definition [ definition - weak - coarse - mv - pair ] . [ definition - coarse - mv - pair ] let @xmath4 be a @xmath2-algebra filtered by @xmath166 and let @xmath6 be a positive number . an @xmath6-controlled mayer - vietoris pair for @xmath4 is a quadruple @xmath826 such that for some positive number @xmath619 . 1 . @xmath628 is a completely coercive decomposition pair for @xmath4 of order @xmath6 with coercitivity @xmath619 . @xmath827 is an @xmath6-controlled @xmath828-neighborhood-@xmath2-algebra for @xmath650 ; 3 . the pair @xmath829 has the cia property for any positive number @xmath143 with coercitivity @xmath619 as defined above . the positive number @xmath619 is called the * coercity * of the @xmath6-controlled mayer - vietoris pair @xmath826 . if @xmath4 is a unital @xmath2-algebra filtered by @xmath166 and if @xmath826 is an @xmath6-controlled mayer - vietoris pair , we will view @xmath834 the unitarization of @xmath835 as @xmath836 and similarly for @xmath837 and @xmath838 . [ example - roe ] let @xmath839 be a proper metric discrete space , let @xmath840 and @xmath841 be subsets in @xmath495 such that @xmath842 and let @xmath6 be a positive number . assume that @xmath843 and @xmath844 , where @xmath845 and @xmath846 are families of @xmath606-disjoint subsets of @xmath495 with uniformly bounded diameter for some positive number @xmath847 . let us consider as in example [ example - roe - algebras ] for @xmath848 the sets @xmath849 of locally compact operators on @xmath798 with support in @xmath850 and let us consider then the subalgebra @xmath851 of @xmath797 of operators with support in @xmath852 for some fixed positive number @xmath143 with @xmath812 . let @xmath853 for @xmath311 integer be the characteristic function of @xmath854 set @xmath855 then @xmath856 is norm decreasing . since @xmath857 for every @xmath597 in @xmath858 , we obtain @xmath859 for every @xmath596 in @xmath860 and @xmath597 in @xmath861 . since @xmath862 lies in @xmath863 , we see that @xmath826 is an @xmath6-controlled mayer - vietoris pair with coercitivity @xmath13 . next lemma makes in the set - up of controlled mayer - vietoris pairs the connection between the decompositions of proposition [ proposition - decom - unitaries ] corresponding respectively to an @xmath40-@xmath6-unitary and its adjoint . [ proposition - decom - unitaries - adjoint ] for every positive number @xmath619 , there exists a control pair @xmath554 such that the following holds . let @xmath4 be any unital filtered @xmath2-algebra , let @xmath6 be any positive number , let @xmath864 be any @xmath6-controlled mayer - vietoris pair for @xmath4 at order @xmath6 with coercitivity @xmath619 , and let @xmath40 and @xmath130 be positive numbers with @xmath865 and @xmath9 . assume that for some @xmath40-@xmath6-unitary @xmath42 in some @xmath866 , there exist @xmath867 and @xmath868 two @xmath40-@xmath130-unitaries in @xmath869 and @xmath870 and @xmath871 two @xmath40-@xmath130-unitaries in @xmath872 such that @xmath873 and @xmath874 . then there exists an integer @xmath86 and @xmath875 and @xmath876 respectively @xmath632-@xmath877-unitaries in @xmath878 and @xmath879 such that * @xmath880 ; * @xmath881 is homotopic to @xmath882 as an @xmath632-@xmath877-unitary in @xmath883 for @xmath650 . moreover , if @xmath884 and @xmath885 lie in @xmath886 for @xmath650 then @xmath887 and @xmath888 can be chosen such that @xmath889 lies in @xmath890 for @xmath650 . let @xmath891 be control pair as in proposition [ proposition - decom - unitaries ] . since @xmath892 and @xmath893 are for @xmath848 homotopic to @xmath160 as a @xmath894-@xmath143-unitary of @xmath158 for every positive number @xmath143 ( * ? ? * lemma 1.20 ) , then up to replacing @xmath895 by @xmath896 , there exists an integer @xmath86 and @xmath745 and @xmath746 be two @xmath632-@xmath897-unitaries respectively in @xmath898 and @xmath899 such that if we set @xmath900 for @xmath848 , then * @xmath901 ; * @xmath902 is in @xmath903 for @xmath848 . * @xmath904 is homotopic to @xmath905 as an @xmath632-@xmath897-unitaries in @xmath903 for @xmath848 . then @xmath906 and hence we have @xmath907 since @xmath908 for @xmath848 and in view of cia property , there exist @xmath57 in @xmath909 with propagation less than @xmath910 such that * @xmath911 is in @xmath912 ; * @xmath913 ; * @xmath914 . in particular some control pair @xmath915 depending only on @xmath619 . * @xmath57 is an @xmath916-@xmath917-@xmath130-unitary in @xmath918 ; * @xmath57 is homotopic to @xmath919 as an @xmath916-@xmath917-@xmath130-unitary in @xmath920 ; * @xmath57 is homotopic to @xmath921 as an @xmath916-@xmath917-@xmath130-unitary in @xmath922 . let us set @xmath923 and @xmath924 . then @xmath887 and @xmath888 satisfy the required properties for some suitable control pair depending only on @xmath619 . in this section , we discuss the example of a controlled mayer - vietoris pair associated to groupods . our method in this paper provides a different approach to the controlled mayer - vietoris sequence in the context of crossed product @xmath1-algebras in @xcite . recall first the definition of a proper symmetric length on an tale groupod . let @xmath280 be an tale groupod , with compact base space @xmath668 . a proper symmetric length on @xmath280 is a continuous proper map @xmath925 such that * @xmath926 if and only @xmath450 is a unit of @xmath280 ; * @xmath927 for any @xmath450 in @xmath280 ; * @xmath928 for any @xmath450 and @xmath453 in @xmath280 composable . let @xmath280 be an tale groupod with compact base space @xmath668 and source map and range map @xmath929 equipped with and symmetric proper length @xmath402 . then , if we set @xmath930then the reduced @xmath2-algebra @xmath931 of @xmath280 is filtered by @xmath932 with @xmath933 for all positive number @xmath6 . for every open subset @xmath934 of @xmath668 and every positive number @xmath6 , set @xmath935 then @xmath936 is an open subset of @xmath668 and @xmath937 . if @xmath938 and @xmath507 are subsets of @xmath939 , then we set @xmath940 and @xmath941 . for every open subset @xmath934 of @xmath668 and every positive number @xmath6 , let @xmath942 be the subgroupod of @xmath943 generated by @xmath944 . then @xmath945 is an open subgroupod of @xmath280 . let us set @xmath946then @xmath947 is a closed linear subspace of @xmath948 and for every positive number @xmath606 and @xmath949 with @xmath950 , then we have @xmath951 . in particular , if @xmath952 , then @xmath953 is a @xmath6-controlled @xmath947-neighborhood-@xmath2-algebra . let @xmath954 and @xmath955 be two open subsets of @xmath668 such that @xmath956 . fix @xmath952 . set @xmath957 and @xmath958 . using partition of unity relatively to @xmath954 and @xmath955 , we see that @xmath628 is a completely coercitive decomposition pair of order @xmath6 for @xmath959 with coercitivity @xmath13 . let us set also @xmath960 and @xmath961 let @xmath143 be positive number with @xmath616 , let @xmath40 be a positive number , let @xmath596 be an element of @xmath962 and let @xmath597 be an element of @xmath963 such that @xmath964 . let @xmath965 and @xmath966 be respectively elements in @xmath967 and @xmath968 such that @xmath969 and @xmath970 . let @xmath0 be a compact subset of @xmath971 such that all coefficients of @xmath972 have support in @xmath973 and let @xmath974 $ ] be a continuous function with support in @xmath971 and such that @xmath975 for all @xmath365 in @xmath0 . the schur multiplication by @xmath976 @xmath977 extends to a completely positive map @xmath978 of complete norm less than @xmath13 and such that @xmath979 and @xmath980 belongs to @xmath981 . moreover @xmath982 hence , @xmath983 is for every @xmath952 a @xmath6-controlled weak mayer - vietoris pair for @xmath984 with coercitivity @xmath985 . assume that there exists a positive number @xmath615 such that for every compact subset @xmath0 of @xmath955 , there exists a continuous function @xmath974 $ ] with support in @xmath955 that satisfies the following : * @xmath975 for all @xmath365 in @xmath0 ; * the schur multiplication by @xmath976 extends to a completely bounded map @xmath986 with complete norm bounded by @xmath615 and such that for any @xmath18 in @xmath987 with support in @xmath988 , then @xmath989 , then @xmath983 is for every @xmath952 a @xmath6-controlled mayer - vietoris pair for @xmath984 with coercitivity @xmath619 depending only on @xmath615 . in this section , we establish for a control mayer - vietoris pair asssocied to a filtered @xmath2-algebra @xmath4 a controlled exact sequence that allows to compute quantitative @xmath0-theory of @xmath4 up to a certain order . let @xmath4 be a unital @xmath2-algebra filtered by @xmath7 , let @xmath6 be a positive number and let @xmath826 be a @xmath6-controlled mayer - vietoris pair for @xmath4 . we denote by @xmath990 and @xmath991 the obvious inclusion maps . [ prop - half - exact]for every positive number @xmath619 , there exists a control pair @xmath554 such that for any filtered @xmath2-algebra @xmath4 , any positive number @xmath6 and any @xmath6-controlled weak mayer - vietoris pair @xmath826 for @xmath4 with coercitiviy @xmath619 , then the composition @xmath992 is @xmath554-exact at order @xmath6 . let us first assume that @xmath4 is unital . in the even case let @xmath993 and @xmath994 be respectively element in @xmath995 and @xmath996 such that @xmath997 in @xmath145 . in view of lemma [ lemma - almost - canonical - form ] , we can up to rescaling @xmath40 assume without loss of generality there exist integer @xmath998 and @xmath5 with @xmath999 and two @xmath40-@xmath6-projections @xmath84 and @xmath1000 in @xmath866 such that * @xmath1001 is an element in the matrix algebra @xmath1002 ; * @xmath1003 is an element in the matrix algebra @xmath1004 ; * @xmath1005_{{\varepsilon},s}$ ] ; * @xmath1006_{{\varepsilon},s}$ ] . up to stabilization , we can also assume that @xmath84 and @xmath1000 are homotopic as @xmath40-@xmath143-projections in @xmath866 . let @xmath81 be the control pair of proposition [ prop - conjugate ] . up to stabilisation there exists @xmath42 a @xmath88-@xmath1007-unitary in @xmath866 such that @xmath1008 . up to replacing @xmath42 by @xmath69 , @xmath84 by @xmath1009 and @xmath1000 by @xmath1010 , we can assume in view of lemma [ cor - example - homotopy ] that @xmath42 is homotopic to @xmath160 as a @xmath1011-@xmath1012-unitary in @xmath866 . according to proposition [ proposition - decom - unitaries ] , then for some control pair @xmath1013 depending only on @xmath81 and @xmath619 with @xmath1014 and up to stabilization , there exist @xmath745 and @xmath746 some @xmath766-@xmath1015 unitaries in @xmath866 such that * @xmath748 is an element in the matrix algebra @xmath1016 for @xmath650 ; * @xmath1017 notice that @xmath1018 is an element in the matrix algebra @xmath1019 and @xmath1020 is an element in the matrix algebra @xmath1021 . by definition [ def - cia ] of the cia property , there exists @xmath19 in @xmath1022 such that @xmath1023 and @xmath1024 let us set then @xmath1025 since @xmath1026 and @xmath1027 , up to stabilization , in view of the proof of ( * ? ? ? * lemma 1.9 ) and according to ( * ? ? ? * lemma 1.7 ) , we know that for some control pair @xmath1028 depending only on @xmath81 and @xmath619 and such that @xmath1029 , then for @xmath848 * @xmath1030 is an @xmath1031-@xmath1032-projection in @xmath1033 ; * @xmath1030 is homotopic to @xmath84 as an @xmath1031-@xmath1032-projections in @xmath1033 ; * @xmath46 is connected to @xmath1030 as an @xmath1031-@xmath1032-projections in @xmath1033 ; and * @xmath1034 is an @xmath1031-@xmath1032-projection in @xmath1035 ; * @xmath1036 is homotopic to @xmath1000 as an @xmath1031-@xmath1032-projections in @xmath1035 ; * @xmath46 is connected to @xmath1037 as an @xmath1031-@xmath1032-projections in @xmath1035 . now if we set @xmath1038_{\alpha'{\varepsilon},l'_{\varepsilon}s}$ ] in @xmath1039 , we have that @xmath1040 in @xmath1041 and @xmath1042 in @xmath1043 . a similar proof can be carried out in the odd case but we can also use the controlled bott periodicity ( * ? ? ? * lemma 4.6 ) . the non unital case can be proved in a similar way using lemma [ proposition - decom - unitaries - non - unital ] , noticing that in view of the proof of proposition [ prop - conjugate ] and following the proof of the unital case above above , we can assume that @xmath42 which is now a @xmath1044-@xmath1045-unitary in @xmath153 is such that @xmath1046 has coefficient in @xmath4 ( see the proofs of ( * ? ? ? * lemma 1.11 and corollary 1.31 ) ) . in this subsection , we introduce the quantitative boundary map in the controlled mayer - vietoris sequence for quantitative @xmath0-theory of filtered @xmath1-algebras . [ lemma - technic]for every positive number @xmath619 , there exists a control pair @xmath1047 such that the following holds : let @xmath4 be a unital @xmath2-algebra filtered by @xmath166 , let @xmath6 be a positive number and let @xmath826 be a @xmath6-controlled weak mayer - vietoris pair for @xmath4 with coercitivity @xmath619 . let @xmath40 and @xmath143 be positive numbers with @xmath538 and @xmath1048 , let @xmath998 and @xmath5 be integers and let @xmath42 in @xmath1049 , @xmath57 in @xmath1050 and @xmath1051 be @xmath40-@xmath143-unitaries in @xmath1052 such that * @xmath1053 is an element in the matrix algebra @xmath1054 for @xmath650 ; * @xmath1055 . then , 1 . there exists a @xmath766-@xmath1015-projection @xmath53 in @xmath1056 such that * @xmath1057 is an element in the matrix algebra @xmath1058 ; * @xmath1059 ; * @xmath1060 . 2 . if @xmath53 and @xmath164 are two @xmath766-@xmath1015-projections in @xmath1056 that satisfies the first point , then @xmath1061_{\lambda{\varepsilon},k_{\varepsilon}s}=[q',n]_{\lambda{\varepsilon},k_{\varepsilon}s}$ ] in @xmath1062 . 3 . let @xmath1063 and @xmath1064 be two pairs of @xmath40-@xmath143-unitaries in @xmath1052 satisfying the assumption of the lemma and let @xmath53 and @xmath164 be @xmath766-@xmath1015-projections in @xmath1056 that satisfies the first point relatively to respectively @xmath1063 and @xmath1064 , then @xmath1061_{\lambda{\varepsilon},k_{\varepsilon}s}=[q',n]_{\lambda{\varepsilon},k_{\varepsilon}s}$ ] in @xmath1062 . since @xmath64 is an @xmath40-@xmath143-unitary , we have that @xmath1065 since @xmath1066 , we deduce that @xmath1067 with notations as in definition [ definition - coarse - mv - pair ] , let @xmath19 be an element in @xmath1068 such that @xmath1069 and @xmath1070 and set @xmath1071 then @xmath53 is close to a @xmath1072-@xmath1073-projection and thus we obtain in view of lemma [ lemma - almost - closed ] that there exists a control pair @xmath1047 , depending only on @xmath619 such that the conclusion of the first point is satisfied . with notations as in lemma [ lemma - technic ] and in view of lemma [ lemma - almost - closed ] , if @xmath53 and @xmath164 are @xmath766-@xmath1015-projections of @xmath1056 that satisfies the first point , then @xmath1074_{10\lambda{\varepsilon},k_{\varepsilon}s}=[q',n]_{10\lambda{\varepsilon},k_{\varepsilon}s}.\ ] ] if @xmath1063 and @xmath1064 are two pairs of @xmath40-@xmath143-unitaries in @xmath1052 that satisfy the assumption of the lemma and let @xmath53 and @xmath164 be @xmath766-@xmath1015-projections in @xmath1056 that satisfy the first point relatively to @xmath1063 and @xmath1064 . then @xmath1075 and hence @xmath1076 . hence using the cia condition , we see that there exists @xmath57 in @xmath1077 such that @xmath1078 is in @xmath1079 , @xmath1080 and @xmath1081 . since we have then @xmath1082 and @xmath1083 the last point is consequence of ( * ? ? ? * lemma 1.9 ) and of the second point applied to @xmath1084 and @xmath1085 . [ remark - unicite - lemme - technic]we have a similar statement in the non - unital case with @xmath42 in @xmath1086 and @xmath57 in @xmath1087 such that @xmath1046 and @xmath1088 have coefficients in @xmath4 we are now in position the define the boundary map associated to a controlled mayer - vietoris pair . let @xmath4 be a filtered @xmath2-algebra and let @xmath826 be a @xmath6-controlled weak mayer - vietoris pair for @xmath4 with coercitivity @xmath619 . assume first that @xmath4 is unital . let @xmath554 be a control pair as is proposition [ proposition - decom - unitaries ] . for any positive numbers @xmath40 and @xmath143 with @xmath82 and @xmath1048 and any @xmath40-@xmath143-unitary @xmath42 in @xmath1089 , let @xmath57 be an @xmath40-@xmath143-unitary in some @xmath1090 such that @xmath64 is homotopic to @xmath1091 as a @xmath66-@xmath1073-unitary in @xmath1092 , we can take for instance @xmath1093 ( see lemma [ cor - example - homotopy ] ) . since @xmath1094 as a filtered subalgebra for @xmath650 , then according to proposition [ proposition - decom - unitaries ] and up to replacing @xmath57 by @xmath177 for some integer @xmath86 , there exists @xmath745 and @xmath746 two @xmath1011-@xmath1095-unitaries in @xmath1056 such that * @xmath1053 is an element in the matrix algebra @xmath1096 for @xmath650 ; * for @xmath650 , there exists a homotopy @xmath750}$ ] of @xmath1011-@xmath1097-unitaries between @xmath13 and @xmath674 such that @xmath1098 is an element in the matrix algebra @xmath1099 for all @xmath179 in @xmath180 $ ] . * @xmath1100 . let @xmath1047 be the control pair of lemma [ lemma - technic ] ( recall that @xmath1101 on depends only @xmath619 ) . then if @xmath40 is in @xmath1102 , there exists a @xmath1103-@xmath1104-projection @xmath53 in @xmath1056 such that * @xmath1057 is an element in the matrix algebra @xmath1105 * @xmath1106 ; * @xmath1107 . in view of second point of lemma [ lemma - technic ] , the class @xmath1061_{3\alpha\lambda{\varepsilon},2l_{3{\varepsilon } } k_{3\alpha{\varepsilon } } s}$ ] in @xmath1108 does not depend on the choice of @xmath53 . set then @xmath1109 and @xmath1110 and define @xmath1111_{{\varepsilon},s})=[q , n]_{\alpha_c{\varepsilon},k_c s}$ ] and let us prove that we define in this way a morphism @xmath1112 it is straightforward to check that ( compare with ( * ? ? * chapter 8) ) . * two choices of elements satisfying the conclusion of lemma [ lemma - technic ] relative to @xmath64 give rise to homotopic elements in @xmath1113 ( this is a consequence of lemma [ lemma - technic ] ) . * replacing @xmath42 by @xmath1114 and @xmath57 by @xmath177 gives also rise to the same element of @xmath1115 . applying now proposition [ proposition - decom - unitaries ] to the @xmath6-controlled mayer - vietoris pair @xmath1116,\delta_1),c([0,1 ] , \delta_2),c([0,1],a_{\delta_1}),c([0,1 ] , a_{\delta_2}))\ ] ] for the @xmath2-algebra @xmath1117,a)$ ] filtered by @xmath1118,a_s))_{s>0}$ ] , we see that @xmath1111_{{\varepsilon},s})$ ] * only depends on the class of @xmath42 in @xmath1119 ; * does not depend on the choice of @xmath57 such that @xmath64 is connected to @xmath1120 in @xmath1121 . in the non unital case @xmath1122 is defined similarly by using point ( ii ) of remark [ remark - unicite - lemme - technic ] , noticing that in view of lemma [ lemma - almost - canonical - form - odd ] and up to replacing @xmath40 by @xmath66 , every element @xmath18 in @xmath1123 is the class of a @xmath40-@xmath6-unitary @xmath42 in @xmath153 such that @xmath1046 has coefficients in @xmath4 . it is straightforward to check that @xmath1124 is compatible with the structure isomorphisms . then if we set @xmath1125 then @xmath1126 is a odd degree @xmath1127-controlled morphism of order @xmath6 . let us now define the boundary map in the even case using controlled bott periodicity . for @xmath607 a closed subspace in an @xmath2-algebra , let us define its suspension as @xmath1128 . let @xmath1129 $ ] be the element of @xmath1130 that implements the extension @xmath1131 where @xmath1132 is the evaluation at @xmath1133 . then @xmath1129 $ ] is an invertible element of @xmath1130 and according to ( * ? ? * lemma 1.6 ) , @xmath1134):{\mathcal{k}}_*(b)\to { \mathcal{k}}_*(sb)\ ] ] is a @xmath1135-controlled isomorphism of degree one with controlled inverse @xmath1134^{-1}):{\mathcal{k}}_*(sb)\to { \mathcal{k}}_*(b).\ ] ] let @xmath4 be a @xmath2-algebra filtered by @xmath166 , let @xmath6 be a positive number and let @xmath826 be a @xmath6-controlled weak mayer - vietoris pair for @xmath4 with coercitivity @xmath619 . then @xmath1136 is a @xmath6-controlled weak mayer - vietoris pair for @xmath1137 ( filtered by @xmath1138 ) with coercitivity @xmath619 . set then @xmath1139 and @xmath1140 . let us define in the even case the quantitative boundary map for the @xmath6-controlled mayer - vietoris pair @xmath826 as the @xmath71-controlled morphism of order @xmath6 @xmath1141^{-1})\circ { \mathcal d}^1_{s{\delta}_1,s{\delta}_2,*}\circ{\mathcal{t}}_a([\partial]):{\mathcal{k}}_0(a){{\longrightarrow}}{\mathcal{k}}_1(a_{{\delta}_1}\cap a_{{\delta}_2}).\ ] ] for sake of simplicity , we will rescale @xmath1127 to @xmath71 and use the same control pair in the odd and in the even case for the definition of @xmath1142 as a odd degree @xmath1143-controlled morphism of order @xmath6 . notice that the quantitative boundary map of a @xmath6-controlled weak mayer - vietoris pair is natural in the following sense : let @xmath4 and @xmath232 be filtered @xmath2-algebras , let @xmath826 and @xmath1144 be respectively @xmath6-controlled weak mayer - vietoris pairs for @xmath4 and @xmath232 with coercitivity @xmath619 and let @xmath410 be a filtered morphism such that @xmath1145 and@xmath1146 . then we have @xmath1147 where @xmath1148 is the restriction of @xmath1149 to @xmath832 . in this subsection , we state the mayer - vietoris controlled six term exact sequence associated to a @xmath6-controlled mayer - vietoris of order @xmath6 . [ lemma - bound1 ] there exists a control pair @xmath1013 such that * for any unital filtered @xmath2-algebra @xmath4 filtered by @xmath166 and any subalgebras @xmath362 and @xmath363 of @xmath4 such that @xmath389 and @xmath1150 are respectively filtered by @xmath1151 and @xmath1152 ; * for any positive number @xmath40 with @xmath538 any integers @xmath5 and @xmath998 and any @xmath40-@xmath6-unitaries @xmath1153 in @xmath866 and @xmath1154 in @xmath1155 ; * for any @xmath40-@xmath6-unitaries @xmath867 and @xmath870 respectively in @xmath1156 and @xmath1157 such that * * @xmath1158 ; * * there exists an @xmath40-@xmath6-projection @xmath53 in @xmath1056 such that @xmath1057 is in @xmath1159 and @xmath1061_{{\varepsilon},r}=0 $ ] in @xmath1160 . then there exists an integer @xmath86 and @xmath1161-@xmath1162-unitaries @xmath745 and @xmath746 respectively in @xmath1163 and @xmath1164 such that @xmath1165 . moreover , if @xmath884 lies in @xmath1166 for @xmath650 then @xmath745 and @xmath746 can be chosen such that @xmath1167 lies in @xmath1168 for @xmath650 up to replacing @xmath1169 and @xmath870 respectively by @xmath1170 and @xmath1171 for some integer @xmath86 , we can assume that @xmath53 is homotopic to @xmath1172 as an @xmath40-@xmath6-projection in @xmath1173 . according to lemma [ prop - conjugate ] , there exist * a control pair @xmath285 ; * up to stabilization an @xmath88-@xmath1174-unitary @xmath57 in @xmath1175 with @xmath1078 in @xmath1176 such that @xmath1177 up to take a larger control pair @xmath285 , we can assume that @xmath1178 and @xmath1179 and hence even indeed that @xmath1180 and @xmath1181hence , for some control pair @xmath1182 depending only on @xmath285 , there exist @xmath1183-@xmath1184-unitaries @xmath1185 in @xmath1186 , @xmath887 in @xmath1187 , @xmath1188 in @xmath1189 , @xmath888 in @xmath1190 such that @xmath1191 and @xmath1192 . thus , for a control pair @xmath1193 depending only on @xmath1182 we have , @xmath1194hence we deduce that @xmath1195 . [ cor - bound1 ] for any positive number @xmath619 , there exists a control pair @xmath1013 such that * for any filtered @xmath2-algebra @xmath4 ; * for any positive number @xmath6 and any @xmath6-controlled weak mayer - vietoris pair @xmath826 for @xmath4 with coercitivity @xmath619 ; * for any positive numbers @xmath1196 and @xmath130 with @xmath1197 and @xmath1198 then for any @xmath19 in @xmath125 such that @xmath1199 in @xmath1200 , there exist @xmath596 in @xmath1201 and @xmath597 in @xmath1202 such that @xmath1203 let us assume for sake of simplicity that @xmath4 is unital , the non unital being similar ( just extra notation are added ) . let @xmath19 be an element in @xmath1204 such that @xmath1205 in @xmath1200 . let @xmath1206 be the controlled pair of lemma [ lemma - bound1 ] and let @xmath42 be an @xmath40-@xmath6-unitary in some @xmath866 such that @xmath1207_{{\varepsilon},r}$ ] . then according to the definition of @xmath1208 , we see by using lemma [ lemma - bound1 ] that up to replacing @xmath42 by @xmath1114 for some integer @xmath998 , there exists two @xmath1209-@xmath1210-unitaries @xmath745 and @xmath746 respectively in @xmath1211 and @xmath1212 such that @xmath1213 . then @xmath42 is homotopic to @xmath1214 as a @xmath1215-@xmath1216-unitary in @xmath1217 . from this we deduce that @xmath1218_{4{\lambda}{\varepsilon}',2l_{{\varepsilon}'}r'}\\ & = & [ w_1]_{4{\lambda}{\varepsilon}',2l_{{\varepsilon}'}r'}+[w_2]_{4{\lambda}{\varepsilon}',2l_{{\varepsilon}'}r'}\\ & = & \jmath_{{\delta}_1,*}^{4{\lambda}{\varepsilon}',2l_{{\varepsilon}'}r'}(x_1)+\jmath_{{\delta}_2,*}^{4{\lambda}{\varepsilon}',2l_{{\varepsilon}'}r'}(x_2)\end{aligned}\ ] ] with @xmath1219_{4{\lambda}{\varepsilon}',2l_{{\varepsilon}'}r'}$ ] in @xmath1220 and @xmath1221_{4{\lambda}{\varepsilon},2l_{{\varepsilon}'}r'}$ ] in @xmath1222 . as a consequence , we get following controlled exactness result . [ prop - bound1]for any positive number @xmath619 , there exists a control pair @xmath554 such that for any @xmath2-algebra @xmath4 filtered by @xmath166 , any positive number @xmath6 and any @xmath6-controlled weak mayer - vietoris pair @xmath826 for @xmath4 with coercitivity @xmath619 then the composition @xmath1223 is @xmath554-exact at order @xmath6 . [ lemma - technic2]there exists a control pair @xmath71 such that the following holds : * let @xmath4 be a unital @xmath2-algebra filtered by @xmath7 and let @xmath362 and @xmath363 be subalgebras of @xmath4 such that @xmath389 and @xmath1150 are respectively filtered by @xmath1151 and @xmath1152 ; * let @xmath40 and @xmath143 be positive numbers with @xmath1224 ; * let @xmath5 and @xmath542 be positive integers with @xmath1225 and let @xmath46 an @xmath40-@xmath143 projection in @xmath1226 such that @xmath1227 . assume that that * @xmath46 is homotopic to @xmath1172 as an @xmath40-@xmath143-projection in @xmath1228 ; * @xmath46 is homotopic to @xmath1172 as an @xmath40-@xmath143-projection in @xmath1229 . then there exist an integer @xmath555 with @xmath556 , @xmath745 and @xmath746 in @xmath1230 , @xmath42 in @xmath1231 and @xmath57 in @xmath1232 such that * @xmath1233 is an element in @xmath1234 for @xmath650 ; * @xmath1235 and @xmath1236 * for @xmath650 , then @xmath674 is connected to @xmath1237 by a homotopy of @xmath1238-@xmath633-unitaries @xmath750}$ ] in @xmath1239 such that @xmath1240 is in @xmath1234 for all @xmath179 in @xmath180 $ ] . * @xmath1241 . let @xmath1242 be the control pair of proposition [ prop - conjugate ] , then there exist up to stabilization * @xmath745 an @xmath632-@xmath1007-unitary in @xmath1228 ; * @xmath746 an @xmath632-@xmath1007-unitary in @xmath1229 , such that @xmath1243 and @xmath1244 up to replacing @xmath745and @xmath746 respectively by @xmath1245 and @xmath1246 and up to replacing @xmath895 by @xmath1247 , we can assume that @xmath1248 is an element in the matrix algebra @xmath1249 and @xmath1250 is an element in the matrix algebra @xmath1251 . hence there exists a control pair @xmath1252 , depending only on @xmath1242 and that we can choose larger such that @xmath1253 and @xmath1254 up to replacing @xmath1255 and @xmath1252 respectively by @xmath1256 , @xmath1257 , @xmath1258 and @xmath1259 , we can assume that @xmath674 for @xmath650 is connected to @xmath1260 by a homotopy of @xmath632-@xmath1015-unitaries @xmath750}$ ] in @xmath1261 such that @xmath1262 is in @xmath1263 for all @xmath179 in @xmath180 $ ] . equations ( [ equ - proof - half - exacness-1 ] ) and ( [ equ - proof - half - exacness-2 ] ) imply that for a control pair @xmath1264 , depending only on @xmath1252 , there exist @xmath42 and @xmath57 some @xmath1265-@xmath1266-unitaries respectively in @xmath866 and @xmath1267 such that @xmath1268 [ prop - bound2]for every positive number @xmath619 , there exists a control pair @xmath891 such that for any filtered @xmath2-algebra @xmath4 , any positive number @xmath6 and any @xmath6-controlled weak mayer - vietoris pair @xmath1269 for @xmath4 of order @xmath6 with coercitivity @xmath619 , then the composition @xmath1270 is @xmath554-exact at order @xmath6 . as in the previous proposition , let us assume that @xmath4 is unital . let @xmath19 be an element in @xmath1271 such that @xmath1272 in @xmath1273 and @xmath1274 in @xmath1275 . let @xmath46 be an @xmath40-@xmath6-projection in some @xmath1276 and @xmath5 be an integer such that @xmath1277_{{\varepsilon},s}$ ] . in view of lemma [ lemma - almost - canonical - form ] and up to replacing @xmath40 by @xmath56 , we can assume without loss of generality that @xmath1278 and that @xmath1279 up to stabilization , we can also assume that * @xmath46 is homotopic @xmath1172 as an @xmath40-@xmath143-projection in @xmath1280 ; * @xmath46 is homotopic @xmath1172 as an @xmath40-@xmath143-projection in @xmath1281 . let @xmath1242 be a control pair , @xmath555 be an integer with @xmath556 , let @xmath745 and @xmath746 be in @xmath1282 , let @xmath42 be in @xmath1283 and let @xmath57 be in @xmath1284 that satisfy all together the conclusion of lemma [ lemma - technic2 ] . since @xmath1285 , we can up to replacing @xmath1242 by @xmath1286 and in view of lemma [ lemma - almost - closed ] moreover assume that @xmath64 is homotopic to @xmath1287 as an @xmath632-@xmath1288-unitary of @xmath1239 . since @xmath1289 and @xmath1290 and in view of the definition of the quantitative boundary map of a control mayer - vietoris pair , there exists a control pair @xmath1252 depending only on @xmath1242 and @xmath619 such that @xmath1291_{{{\alpha}}''{\varepsilon},k'_{\varepsilon}s})=[p , l]_{{\varepsilon},{{\alpha}}'\alpha_c{\varepsilon},s , k'_{\varepsilon}k_{c,\lambda{\varepsilon } } s}.\]]hence , if we set @xmath1292_{{{\alpha}}'{\varepsilon},k'_{\varepsilon}s}$ ] , we get @xmath1293 collecting together propositions [ prop - half - exact ] , [ prop - bound1 ] and [ prop - bound2 ] and using naturality of quantitative bott isomorphism , we obtain the controlled six terms exact sequence ( at order @xmath6 ) associated to a weak @xmath6-controlled mayer - vietoris sequence . [ thm - six - term - exact - sequence ] for every positive number @xmath619 , there exists a control pair @xmath71 such that for any @xmath2-algebra @xmath4 filtered by @xmath166 , any positive number @xmath6 and any @xmath6-controlled weak mayer - vietoris pair @xmath1269 for @xmath4 of order @xmath6 , then we have a @xmath71-exact six term exact sequence of order @xmath6 : @xmath1294 in numerous case , the proof of the baum - connes conjecture and of its generalization amounts to show that the @xmath0-theory of some obstruction algebra vanishes @xcite . in this subsection , we apply the controlled mayer - vietoris six - term exact sequence to this situation . let @xmath4 be a filtered @xmath2-algebra . @xmath4 is called * quantitatively @xmath0-contractible * if there exists a positive number @xmath1295 that satisfies the following holds : for any positive numbers @xmath40 and @xmath6 with @xmath1296 , there exists a positive number @xmath130 with @xmath203 such that @xmath1297 is vanishing ( we say that @xmath4 is @xmath0-contractible with * rescaling * @xmath200 ) . 1 . recall that a separable @xmath2-algebre @xmath232 is @xmath0-contractible if the class of the identity map @xmath1298 vanishes in @xmath1299 . according to theorem [ thm - tensor ] , if @xmath232 is @xmath0-contractible , then @xmath357 is quantitatively @xmath0-contractible for any filtered @xmath2-algebra @xmath4 . moreover , the rescaling does not depend on @xmath4 or on @xmath232 ; 2 . let @xmath393 be a finitely generated group and let @xmath4 be a @xmath2-algebra provided with an action of @xmath393 by automorphisms . assume that * the group @xmath393 satisfies the baum - connes conjecture with coefficients . * for any finite subgroup @xmath1300 of @xmath393 the @xmath1301 . then @xmath1302 is quantitatively @xmath0-contractible and the rescaling does not depend on @xmath393 or on @xmath4 . [ remark - universal - rescaling ] it can be proved that there exists a universal rescaling for quantitative @xmath0-contractibility , i.e there exists a positive number @xmath301 with @xmath302 such that every quantitatively @xmath0-contractible @xmath2-algebra is indeed quantitatively @xmath0-contractible with rescaling @xmath301 . [ thm - quant - k - contractible ] let @xmath4 be a filtered @xmath2-algebra . assume that there exists positive numbers @xmath200 and @xmath619 , with @xmath1295 such that for every positive number @xmath6 there exists an @xmath6-controlled mayer - vietoris pair @xmath864 with coercitivity @xmath619 and with @xmath1303 and @xmath832 quantitatively @xmath0-contractible with rescaling @xmath200 . then there exists a positive number @xmath1304 depending only on @xmath200 and on @xmath619 such that @xmath4 is quantitatively @xmath0-contractible with rescaling @xmath1304 . let @xmath1206 be the controlled pair of corollary [ cor - bound1 ] and set @xmath1305 , let @xmath40 and @xmath6 be positive numbers with @xmath1306 and let @xmath19 be an element in @xmath1307 . let @xmath864 be a @xmath6-controlled mayer - vietoris pair for @xmath4 with coercitivity @xmath619 . let @xmath130 be a positive number with @xmath1308 such that @xmath1309 in @xmath1310 for all @xmath365 in @xmath1311 . since @xmath1312 in @xmath1311 and according to corollary [ cor - bound1 ] , then if we set @xmath1313 , there exist an element @xmath596 in @xmath1314 and an element @xmath597 in @xmath1315 such that @xmath1316 let @xmath1317 be a positive number with @xmath1318 such that for @xmath650 , @xmath1319 in @xmath1320 for all @xmath365 in @xmath1321 . then we eventually obtain that @xmath1322 hence @xmath4 is quantitatively @xmath0-contractible with rescaling @xmath1304 . in this section , we formulate a quantitative knneth formula for filtered @xmath2-algebras and we discuss its connection with controlled mayer - vietoris pairs . we also show that finitely generated groups for which the baum - connes conjecture with coefficient holds provides numerous examples of filtered @xmath2-algebras that satisfy the quantitative knneth formula recall that if @xmath4 and @xmath232 are @xmath2-algebras , then there is a morphism @xmath1323 given by the external kasparov product i.e. , @xmath1324 for all @xmath18 in @xmath197 and @xmath19 in @xmath1325 . indeed , in the case of unital @xmath2-algebras , if @xmath46 and @xmath53 are respectively projections in @xmath866 and @xmath1326 and if @xmath42 and @xmath57 are respectively unitary elements in @xmath866 and @xmath1326 , then @xmath1327{{\otimes}}[q])&=&[p{{\otimes}}q];\\ \omega_{a , b , * } ( [ u]{{\otimes}}[q])&=&[u{{\otimes}}q+i_n{{\otimes}}(i_k - q)];\\ \omega_{a , b , * } ( [ p]{{\otimes}}[v])&=&[p{{\otimes}}v+(i_n - p){{\otimes}}i_k ] .\\ \end{aligned}\ ] ] let @xmath4 be a @xmath2-algebra filtered by @xmath7 and let @xmath232 be a @xmath2-algebra ( with a trivial filtration ) . recall that @xmath357 is then filtered by @xmath1328 . let us consider then the quantitative object @xmath1329 . with notations of theorem [ thm - tensor ] , define the @xmath368-control morphism @xmath1330 by @xmath1331 then @xmath1332 induces @xmath1333 in @xmath0-theory , i.e. , @xmath1334 for every positive numbers @xmath6 and @xmath40 with @xmath1335 . [ remark - kunneth ] let @xmath4 be a unital filtered @xmath2-algebra and let @xmath232 be a unital @xmath2algebra . let @xmath40 and @xmath6 be positive numbers with @xmath1336 . 1 . for any @xmath40-@xmath6-projection @xmath46 in some @xmath866 , any integer @xmath87 and any projection @xmath53 in some @xmath1326 then @xmath1337_{{\varepsilon},r}{{\otimes}}[q])=[p{{\otimes}}q+i_l{{\otimes}}(i_k - q),lk]_{{{\alpha}}_{\mathcal{t}}{\varepsilon},h_{{\mathcal{t}},{\varepsilon}}r}$ ] in @xmath1338 . 2 . for any @xmath40-@xmath6-unitary @xmath42 in some @xmath866 and any projection @xmath53 in some @xmath1326 then @xmath1339_{{\varepsilon},r}{{\otimes}}[q])=[u{{\otimes}}q+i_n{{\otimes}}(i_k - q)]_{{{\alpha}}_{\mathcal{t}}{\varepsilon},h_{{\mathcal{t}},{\varepsilon}}r}$ ] in @xmath1340 . the quantitative morphism @xmath1341 is compatible with the kasparov tensorisation ( controlled ) morphsim . [ lemma - compatibility - kasparov ] there exists a control pair @xmath1242 such that the following assertion holds : for any filtered @xmath2-algebra @xmath4 , for any separable @xmath2-algebras @xmath1342 and @xmath436 , any @xmath365 in @xmath1343 and any @xmath371 in @xmath437 , then the following diagram is @xmath1242-commutative . @xmath1344 where @xmath1345 is right multiplication by @xmath371 . let @xmath19 be an element of @xmath1346 . according to point ( vi ) of theorem [ thm - tensor ] and to theorem [ thm - product - tensor ] , there exists a control pair @xmath261 such that @xmath1347 since external kasparov is commutative , we have @xmath1348 using once again theorem [ thm - product - tensor ] and up to rescaling the control pair @xmath261 , we get that @xmath1349 and hence the diagram is commutative . let @xmath4 be a filtered @xmath2-algebra and let @xmath301 be a positive number with @xmath302 . we say that @xmath4 satisfies the quantitative knneth formula with rescaling @xmath301 if @xmath1350 is a quantitative isomorphism with rescaling @xmath301 for every @xmath2-algebra @xmath232 such that @xmath1325 is free abelian group . if a filtered @xmath2-algebra @xmath4 satisfies the quantitative knneth formula , then according to equation ( [ eq - induce - kunneth ] ) @xmath1323 is an isomorphism for every @xmath2-algebra @xmath232 such that @xmath1325 is free . using geometric resolutions , it was proved in @xcite that in this case , the @xmath2-algebra @xmath4 satisfies the knneth formula , i.e. , for any @xmath2-algebra @xmath232 we have a natural short exact sequence @xmath1351 ( see also @xcite for the straightforward generalization to the non nuclear case ) . next theorem provides many examples of filtered @xmath2-algebras that satisfy the quantitative knneth formula and will be proved in section [ subsec - kunneth - bc ] . [ thm - bc - qkunneth ] let @xmath393 be a finitely generated group , let @xmath4 be a @xmath393-@xmath2-algebra . assume that * @xmath393 satisfies the baum - connes conjecture with coefficients . * for each subgroup subgroup @xmath0 of @xmath393 , then @xmath1352 satisfies the knneth formula . then @xmath1353 satisfies the quantitative knneth formula , i.e. , for any @xmath2-algebra @xmath232 such that @xmath1325 is a free abelian group , @xmath1354 is a quantitative isomorphism with rescaling that does not depend on @xmath393 or on @xmath4 . moreover , under above assumption , when the @xmath2-algebra @xmath4 runs through family of @xmath393-@xmath2-algebras and @xmath232 runs through @xmath2-algebras such that @xmath1325 is a free abelian group , the family of quantitative isomorphisms @xmath1355 is uniform . the proof of this two results relies indeed on the quantitative statements of theorem [ thm - quant - surj ] which hold for groups that satisfy the baum - connes conjecture with coefficients . similarly , using the geometric quantitative statements of section [ subsection - quantitative - assembly - map ] and theorem [ thm - geo - statement ] , we can prove the following result : let @xmath495 be a discrete proper metric space with bounded geometry that coarsely embeds into a hilbert space . if @xmath4 satisfies the knneth formula , then @xmath1356 satisfies the quantitative knneth formula with rescaling that does not depend on @xmath495 or on @xmath4 . as for quantitative @xmath0-contractibility ( see remark [ remark - universal - rescaling ] ) , it can be proved that there exists a universal rescaling for the quantitative knneth formula . in this subsection , we state a permanence result ( that we shall prove in next subsection ) for the quantitative knneth formula with respect to controlled mayer - vietoris pairs which satisfy a nuclear type condition . [ definition - nuclear - coarse - mv - pair ] let @xmath4 be a @xmath2-algebra filtered by @xmath166 and let @xmath6 be a positive number . an @xmath6-controlled nuclear mayer - vietoris pair is a quadruple @xmath826 , where @xmath634 and @xmath635 are closed linear subspaces of @xmath10 stable under involution and @xmath1357 and @xmath1358 are respectively @xmath6-controlled @xmath1359 and @xmath635-neighborhood-@xmath2-algebras such that for some positive number @xmath619 and for any @xmath2-algebra @xmath232 1 . @xmath1360 is a coercive decomposition pair for @xmath357 of order @xmath6 with coercitivity @xmath619 . 2 . the pair @xmath1361 has the cia property with coercitivity @xmath619 for any positive number @xmath143 . the positive number @xmath619 is called the * coercity * of the @xmath6-controlled nuclear mayer - vietoris pair @xmath826 . 1 . notice that @xmath1362 and @xmath1363 are respectively @xmath6-controlled @xmath1364 and @xmath1365-neighborhood-@xmath2-algebras . 2 . condition ( ii ) amounts to the following : for any positive numbers @xmath40 and @xmath143 , any @xmath18 in @xmath1366 and any @xmath19 in @xmath1367 such that @xmath819 , then there exists @xmath365 in @xmath1368 satisfying @xmath821 replacing @xmath158 with @xmath232 , we see that example [ example - roe ] and examples of section [ subsection - groupoid ] are indeed @xmath6-controlled nuclear mayer - vietoris pairs ( with the same coercitivity ) . next lemma shows that the controlled boundary maps of a @xmath6-controlled nuclear mayer - vietoris pair are indeed compatible with kasparov external product . [ lemma - mv - kunneth ] for any positive number @xmath619 , there exists a control pair @xmath300 such that the following is satisfied : let @xmath826 be an @xmath6-controlled nuclear mayer - vietoris pair with coercitivty @xmath619 , let @xmath232 and @xmath423 be two @xmath2-algebras and @xmath365 be an element in @xmath1369 , then @xmath1370 we first deal with the case @xmath365 even . according to ( * ? ? ? * lemma 1.6.9 ) , there exists a @xmath2-algebra @xmath382 and homomorphisms @xmath1371 and @xmath1372 such that * the element @xmath1373 $ ] of @xmath1374 induced by @xmath1375 is invertible . * @xmath1376^{-1})$ ] . let @xmath1377 and @xmath1378 be the homorphisms induced by @xmath1375 on tensor products and define similarly @xmath1379 and @xmath1380 the homomorphisms induced by @xmath1381 . by naturality of quantitative boundary morphism of @xmath6-controled mayer - vietoris pairs ( see equation ( [ equ - boundary - natural ] ) of section [ subs - section - boundary ] ) , we get that @xmath1382 according to theorem [ thm - tensor ] , the control morphisms @xmath1383 and @xmath1384 are @xmath1135-invertible with @xmath1135-inverses respectively given by @xmath1385^{-1})$ ] and @xmath1386^{-1})$ ] and hence , there exists a control pair @xmath1242 depending only on @xmath619 and @xmath1135 such that @xmath1387^{-1 } ) \stackrel{({{\alpha}},k)}{\sim } { \mathcal{t}}_{a_{\delta_1}\cap a_{\delta_2},*}([\theta]^{-1})\circ { \mathcal d}_{{\delta}_1{{\otimes}}d,{\delta}_2{{\otimes}}d}.\ ] ] then , using the bifunctoriality of @xmath1388 ( see theorem [ thm - tensor ] ) , we obtain @xmath1389^{-1})\circ \eta_{a,*}\\ & \stackrel{({{\alpha}},k)}{\sim } & { \mathcal{t}}_{a_{\delta_1}\cap a_{\delta_2},*}([\theta]^{-1})\circ { \mathcal d}_{{\delta}_1{{\otimes}}d,{\delta}_2{{\otimes}}d}\circ \eta_{a,*}\\ & \stackrel{({{\alpha}},k)}{\sim } & { \mathcal{t}}_{a_{\delta_1}\cap a_{\delta_2},*}([\theta]^{-1})\circ \eta_{a_{\delta_1}\cap a_{\delta_2},*}\circ { \mathcal d}_{{\delta}_1{{\otimes}}b,{\delta}_2{{\otimes}}b}\\ & \stackrel{({{\alpha}},k)}{\sim } & { \mathcal{t}}_{a_{\delta_1}\cap a_{\delta_2},*}(z)\circ { \mathcal d}_{{\delta}_1{{\otimes}}b,{\delta}_2{{\otimes}}b},\end{aligned}\ ] ] where the third line is once again the consequence of naturality of quantitative boundary morphism of @xmath6-controled mayer - vietoris pairs ( see equation ( [ equ - boundary - natural ] ) of section [ subs - section - boundary ] ) . if @xmath365 is an odd element , recall that @xmath1129 $ ] is the invertible element of @xmath1390 implementing the boundary morphism of the extension @xmath1391\stackrel{ev_0}{{{\longrightarrow } } } { \mathbb c}{{\longrightarrow}}0.\ ] ] then there exists a control pair @xmath1242 depending only on @xmath619 and on the control pairs of theorems [ thm - tensor ] and [ thm - product - tensor ] such that @xmath1392^{-1})\circ { \mathcal d}^1_{s{\delta}_1{{\otimes}}b , s{\delta}_2{{\otimes}}b,*}\circ { \mathcal{t}}_{a{{\otimes}}b , * } ( [ \partial])\\ & \stackrel{({{\alpha}},k)}{\sim}{\mathcal{t}}_{a_{\delta_1}\cap a_{\delta_2},*}(z)\circ { \mathcal{t}}_{(a_{\delta_1}\cap a_{\delta_2}),*}(\tau_b([\partial]^{-1}))\circ { \mathcal d}^1_{s{\delta}_1{{\otimes}}b , s{\delta}_2{{\otimes}}b,*}\circ { \mathcal{t}}_{a{{\otimes}}b , * } ( [ \partial])\\ & \stackrel{({{\alpha}},k)}{\sim}{\mathcal{t}}_{a_{\delta_1}\cap a_{\delta_2},*}(\tau_b([\partial]^{-1}){{\otimes}}_bz)\circ { \mathcal d}^1_{s{\delta}_1{{\otimes}}b , s{\delta}_2{{\otimes}}b,*}\circ { \mathcal{t}}_{a{{\otimes}}b , * } ( [ \partial])\\ & \stackrel{({{\alpha}},k)}{\sim } { \mathcal d}^1_{s{\delta}_1{{\otimes}}b',s{\delta}_2{{\otimes}}b',*}\circ { \mathcal{t}}_{a,*}(\tau_b([\partial]^{-1}){{\otimes}}_bz)\circ { \mathcal{t}}_{a,*}(\tau_{b,*}([\partial]))\\ & \stackrel{({{\alpha}},k)}{\sim } { \mathcal d}^1_{s{\delta}_1{{\otimes}}b',s{\delta}_2{{\otimes}}b ' , * } \circ { \mathcal{t}}_{a,*}(z),\end{split}\ ] ] where * the first @xmath1393 holds by definition of @xmath1394 ; * the @xmath1393 is a consequence of point ( vi ) of theorem [ thm - tensor ] ; * the third and fifth @xmath1393 are consequences of theorem [ thm - product - tensor ] ; * the fourth @xmath1393 is a consequence of the even case ; similarly , we can prove that there exists a control pair @xmath1242 depending only on @xmath619 and on the control pair of theorems [ thm - tensor ] and [ thm - product - tensor ] such that @xmath1395 and hence we obtain the result in the odd case . [ thm - mv - kunneth ] let @xmath4 be a filtered @xmath2-algebra . assume there exists positive numbers @xmath619 and @xmath200 with @xmath302 that satisfies the following : for any positive number @xmath6 , there exists an @xmath6-controlled nuclear mayer - vietoris pair @xmath864 with coercitivity @xmath619 such that @xmath1303 and @xmath832 satisfies the quantitative knneth formula with rescaling @xmath301 . then @xmath4 satisfies the quantitative knneth formula with rescaling @xmath1396 for a positive number @xmath1396 with @xmath1397 depending only on @xmath619 and @xmath301 . the proof of this theorem will be given in next subsection . the idea of the proof of theorem [ thm - mv - kunneth ] is to use the controlled exactness persistence properties stated in corollary [ cor - bound1 ] and lemma [ lemma - technic2 ] to carry out a controlled five lemma argument . this requires some preliminary work in order to formulate the quantitative knneth formula in terms of an ( even degree ) controlled morphism between quantitative @xmath0-theory groups . we denote by @xmath1398 the suspension algebra of a @xmath2-algebra @xmath232 . let @xmath4 be a filtered @xmath2-algebra , let @xmath232 be a @xmath2-algebra and let @xmath1129 $ ] be the invertible element in @xmath1390 that implements the boundary of the bott extension , @xmath1399 . applying lemma [ lemma - compatibility - kasparov ] to @xmath1400 and @xmath1401)$ ] , we see that we only have to consider the odd case @xmath1402 let @xmath1403 be the two torus . let us view @xmath1404 as an open subset of @xmath1405 via the inclusion map @xmath1406 . [ lemma - eight ] possibly rescaling the control pair @xmath1407 , then for any filtered @xmath2-algebra @xmath4 , the filtered and semi - split extension of filtered @xmath2-algebras @xmath1408 has a vanishing controlled boundary map @xmath1409 by using controlled bott periodicity ( * ? ? ? * lemma 4.6 ) and in view of ( * ? ? ? * proposition 3.19 ) , we only need to check that the result holds for @xmath1410 but @xmath1411 let @xmath1412 be an @xmath40-@xmath6-unitary in @xmath1413 and define then @xmath1414then @xmath1415 is a @xmath147-@xmath1416-unitary in @xmath1417 . moreover , under the identification of equation ( [ equation - eight ] ) , we have @xmath1418 and hence @xmath1419 then the result holds by construction of the controlled boundary map in the odd case . in consequence , by using the controlled six term exact sequence ( theorem [ thm - six term ] ) associated to the semi - split and filtered extension of equation ( [ equation - extension - eight ] ) , then if we set @xmath1420 we obtain the following corollary . [ corollary - eight ] there exists a controlled pair @xmath261 such that for any filtered @xmath2-algebra @xmath4 , then @xmath1421 is a @xmath261-controlled isomorphism . notice that , by construction of the controlled boundary map associated to controlled mayer - vietoris pair @xmath864 , we have that @xmath14220,1[^2,a_{{\delta}_1}),c_0({{\mathbb t}}^2\setminus ] 0,1[^2,a_{{\delta}_2}),*}\circ q_{a{{\otimes}}b,*}.\ ] ] let us set with notation of corollary [ corollary - eight ] @xmath1423 set @xmath1424{{\otimes}}\tau_{c_0(0,1)}([\partial])$ ] and define then the @xmath1425-controlled morphism @xmath1426 since @xmath365 is an invertible element in @xmath1427 and hence , according to theorem [ thm - product - tensor ] and to corollary [ corollary - eight ] , * if @xmath1332 is a quantitative isomorphism with rescaling @xmath301 then there exists a positive number @xmath1396 depending only on @xmath301 with @xmath1397 such that @xmath1428 is a quantitative isomorphism with rescaling @xmath1396 ; * if @xmath1428 is a quantitative isomorphism with rescaling @xmath1396 , then there exists a positive number @xmath301 depending only on @xmath1396 with @xmath302 such that @xmath1332 is a quantitative isomorphism with rescaling @xmath301 . according to lemma [ lemma - compatibility - kasparov ] , we get that @xmath1429){{\otimes}}\tau_b([\partial])\]]for some control pair @xmath1101 that does not depend on @xmath4 or on @xmath232 . let us set @xmath1430 and denote for positive numbers @xmath1431 and @xmath130 with @xmath132 and @xmath9 respectively by @xmath1432 and @xmath1433 the structure morphisms corresponding to @xmath1434 and @xmath1435 . for @xmath1436 , let us define the @xmath1437-controlled morphism @xmath1438 by @xmath1439notice that @xmath1440 is obviously natural with respect to filtered morphisms . since @xmath1129 $ ] is invertible in @xmath1390 and according to equation ( [ equation - bott - kunneth ] ) and theorem [ thm - product - tensor ] , theorem [ thm - mv - kunneth ] is equivalent to the following statement : let @xmath4 be a filtered @xmath2-algebra . assume there exists positive numbers @xmath619 and @xmath200 with @xmath302 that satisfies the following : for any positive number @xmath6 , there exists an @xmath6-controlled nuclear mayer - vietoris pair @xmath864 with coercitivity @xmath619 such that for any @xmath2-algebra @xmath232 with @xmath1325 free abelian , then @xmath1441 and @xmath1442 are quantitative isomorphisms with rescaling @xmath301 . then for any for any @xmath2-algebra @xmath232 with @xmath1325 free abelian , @xmath1443 is a quantitative isomorphism with coercity @xmath1396 for a positive number @xmath1396 with @xmath1397 depending only on @xmath301 and @xmath619 . notice that , by controlled bott periodicity , we only need to consider the odd case . let us introduce some notations that we will use throughout the proof . let @xmath57 be a unitary in a unital @xmath2-algebra @xmath232 . define the unitary of @xmath1444 @xmath1445{{\longrightarrow}}u_2(b);\,t\mapsto \left(\begin{smallmatrix}v&0\\ 0&1\end{smallmatrix}\right)\cdot \left(\begin{smallmatrix}\cos \frac{t\pi}{2}&\sin\frac{t\pi}{2}\\ -\sin \frac{t\pi}{2}&\cos\frac{t\pi}{2}\end{smallmatrix}\right)\cdot \left(\begin{smallmatrix}v^*&0\\ 0&1\end{smallmatrix}\right)\cdot \left(\begin{smallmatrix}\cos\frac{t\pi}{2}&-\sin \frac{t\pi}{2}\\ \sin\frac{t\pi}{2}&\cos \frac{t\pi}{2}\end{smallmatrix}\right),\ ] ] and the projection of @xmath1444 @xmath1446{{\longrightarrow}}m_2(b ) ; \ , t\mapsto r_v(t)\cdot \left(\begin{smallmatrix}1&0\\ 0 & 0\end{smallmatrix}\right ) r^*_{v}(t).\ ] ] define also @xmath1447\times [ 0,1]\to m_2({\mathbb c});\ , ( s , t ) \mapsto r_{e^{2\imath \pi s}}(t)\cdot \left(\begin{smallmatrix}1&0\\ 0 & 0\end{smallmatrix}\right ) r^*_{e^{2\imath \pi s}}(t).\ ] ] then , if we view the @xmath1448-sphere @xmath1449 as the one point compactification of @xmath1404 , then @xmath1450 is a rank one projection in @xmath1451 and @xmath1452-[1]$ ] is a generator for @xmath1453 . for a @xmath6-controlled nuclear mayer - vietoris pair @xmath864 with respect to a filtered @xmath2-algebra @xmath4 , let @xmath1454 and @xmath1455 be the morphisms respectively induced by the inclusions of @xmath2-algebras @xmath1456 , @xmath1457 , @xmath1458 and @xmath1459 . in the same way , we define @xmath1460 and @xmath1461 let us now compute @xmath1463 for @xmath4 , a unital filtered @xmath2-algebra , @xmath232 a unital @xmath2-algebra and @xmath18 an element in @xmath1464 . let us consider first the case @xmath1292_{{\varepsilon},r}{{\otimes}}[p]$ ] where @xmath42 is an @xmath40-@xmath6-unitary in some @xmath866 and @xmath46 is a projection in some @xmath1465 . let us set @xmath1466 . then @xmath1467 is an @xmath40-@xmath6-unitary in @xmath1468 . according to remark [ remark - kunneth ] , then @xmath1469_{{\varepsilon},r}{{\otimes}}[p])=j_{a,*}^{{{\alpha}}_{\mathcal{t}}{\varepsilon},h_{k_{\mathcal{t}},{\varepsilon}}r}\circ \tau^{{\varepsilon},r}_a(z)(v_{u , p}).\ ] ] it is well known that @xmath1470-[p_1]$ ] ( with @xmath1471 ) . let us define for @xmath42 and @xmath1472 some @xmath40-@xmath6-unitaries in @xmath866 and @xmath46 a projection in @xmath1465 @xmath1473 then @xmath1474 is an @xmath40-@xmath6-unitary in @xmath1475 . moreover , if @xmath1476 then we see that @xmath1477_{{\varepsilon},r}=0 $ ] in @xmath1478 and hence @xmath1479_{{\varepsilon},r}{{\otimes}}[p])=[w_{u , u^*,p}]_{{{\alpha}}_{\mathcal{f}}{\varepsilon},h_{{\mathcal{f}},{\varepsilon}}r}.\ ] ] consider now the case @xmath1292_{{\varepsilon},r}{{\otimes}}([p]-[p(0)])$ ] , where @xmath42 is an @xmath40-@xmath6-unitary in some @xmath1480 and @xmath46 is a projection in some @xmath1481 . for @xmath40-@xmath6-unitaries @xmath42 and @xmath1472 in @xmath1480 and a projection @xmath46 in @xmath1480 , set @xmath1482 then @xmath1483 is an @xmath40-@xmath6-unitary in @xmath1484 and we can easily check that @xmath1485_{{\varepsilon},r}$ ] is in @xmath1486 for all positive numbers @xmath40 and @xmath6 with @xmath190 . moreover , according to remark [ remark - kunneth ] , we have @xmath1487_{{\varepsilon},r}{{\otimes}}([p]-[i_k]))=[w'_{u , u^*,p}]_{{{\alpha}}_{\mathcal{f}}{\varepsilon},h_{{\mathcal{f}},{\varepsilon}}r}.\ ] ] consider now the case @xmath1488_{{\varepsilon},r}{{\otimes}}[p]$ ] , where @xmath53 is an @xmath40-@xmath6-projection in some @xmath866 , @xmath998 is an integer with @xmath999 and @xmath46 is a projection in some @xmath1326 . let us set when @xmath53 is an @xmath40-@xmath6-projection in some @xmath866 , @xmath998 is a positive number and @xmath1489 and @xmath1490 are projections in some @xmath1326 @xmath1491 with @xmath1492 in @xmath1493 . then @xmath1494 is an @xmath40-@xmath6-projection in @xmath1495 . according to lemma [ lemma - compatibility - kasparov ] , we have that @xmath1496_{{\varepsilon},r}{{\otimes}}([p]{{\otimes}}z)).\ ] ] since @xmath1497{{\otimes}}z=[p{{\otimes}}p_{bott}]-[p{{\otimes}}p_{1}]$ ] in @xmath1498 , we see according to remark [ remark - kunneth ] that @xmath1499_{{\varepsilon},r}{{\otimes}}[p])=[e_{q , m , p{{\otimes}}p_{bott},p{{\otimes}}p_1},2nk]_{{\varepsilon},r}.\ ] ] consider finally the case of @xmath1500_{{\varepsilon},r}{{\otimes}}([p'_1]-[p'_2])$ ] , where @xmath164 is an @xmath40-@xmath6-projection in some @xmath1480 , @xmath1501 is an integer and @xmath1502 and @xmath1503 are projections in some @xmath1504 , with @xmath1505 . then @xmath1506_{{\varepsilon},r}{{\otimes}}([p'_1]-[p'_2]))=[e_{q',m',p'_1,p'_2},nk]_{{\varepsilon},r}.\ ] ] let @xmath864 be an @xmath6-controlled nuclear mayer - vietoris pair with coercitivity @xmath619 such that for any @xmath2-algebra @xmath232 with @xmath1325 free abelian , then @xmath1441 and @xmath1442 are quantitative isomorphisms with rescaling @xmath301 . let us check that there exists a positive number @xmath1396 depending only on @xmath301 and @xmath619 , with @xmath1397 such that for any positive numbers @xmath40 and @xmath143 with @xmath40 in @xmath1508 and @xmath1509 , then @xmath1510 satisfies the @xmath1507-condition of definition [ def - quantitative - morphism ] . we can assume without loss of generality that @xmath4 and @xmath232 are unital . moreover , up to replacing @xmath232 by @xmath1511 , we can assume that there exists a system of generators of @xmath1512 given by classes of projections . let us fix such a system of generator for @xmath1512 and fix also a system of generator for @xmath1513 . as discussed in section [ subsubsection - preliminary].1 , we only need to consider the odd case , i.e show that the control morphism @xmath1514 sastifies the required conditions . according to lemma [ lemma - mv - kunneth ] , there exists a control pair @xmath300 depending only on @xmath619 with @xmath1515 such that @xmath1516 at order @xmath6 . for @xmath40 in @xmath1517 , let @xmath1518 be a positive numbers with @xmath1519 such that @xmath1520 for all @xmath18 in @xmath1521 such that @xmath1522 . [ proposition - technic]there exists a control pair @xmath261 depending only on @xmath301 and @xmath619 such that for any positive numbers @xmath1523 and @xmath130 with @xmath1524 and @xmath1525 , for any @xmath18 in @xmath1526 such that @xmath1527 in @xmath1528 , there exist * an element @xmath1529 in @xmath1530 and an element @xmath1531 in @xmath1532 ; * an integer @xmath5 and an @xmath1533-unitary @xmath1534 in @xmath1535 ; * for @xmath650 , a @xmath1536-unitary @xmath1537 in @xmath1538 with @xmath1539 in @xmath1540 ; such that 1 . @xmath1541 ; 2 . @xmath1542_{{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\varepsilon}}r}=0 $ ] in @xmath1543 ; 3 . @xmath1544_{{\lambda}{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}{\varepsilon}}h_{\varepsilon}r'}$ ] is in @xmath1545 and @xmath1546_{{\lambda}{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}{\varepsilon}}h_{\varepsilon}r'}$ ] for @xmath650 ; 4 . @xmath1547 ; the proof of the proposition is quite long so we split it in several steps . * step 1 : * let @xmath18 be an element in @xmath1548 . then there exists integers @xmath87 and @xmath102 and * for @xmath1549 * * an @xmath40-@xmath143-unitary @xmath1550 in some @xmath1551 ; * * a projection @xmath1552 in some @xmath1553 with @xmath1554 $ ] in the prescribed system of generators for @xmath1512 . * for @xmath1555 * * an @xmath40-@xmath143-unitary @xmath1556 in some @xmath1557 ; * * a projection @xmath1558 in some @xmath1559 with @xmath1560-[{p'_i(0)}]$ ] in the prescribed system of generators for @xmath1561 ; such that @xmath1562_{{\varepsilon},s}{{\otimes}}[p_i]+\sum_{i=1}^{l'}[u'_i]_{{\varepsilon},s}{{\otimes}}([p'_i]-[{p'_i(0)}]).\ ] ] assume that @xmath1563 in @xmath1564 . using morita equivalence , and up to replacing @xmath4 by @xmath1565 and @xmath232 by @xmath1566 with @xmath1567 and @xmath1568 , we can assume that @xmath1569 for @xmath1570 and @xmath1571 for @xmath1572 . using lemma [ lemma - almost - canonical - form - odd ] , we can moreover assume without loss of generality that @xmath1573 for @xmath1572 . let us set @xmath1574 then @xmath1534 is an @xmath40-@xmath143-unitary in @xmath1575 and @xmath1542_{{\varepsilon},s}$ ] is in @xmath1576 . in view of equations ( [ equ - f ] ) and ( [ equ - f - suspension ] ) , we see that @xmath1542_{{{\alpha}}_{{\mathcal{f}}}{\varepsilon},k_{{\mathcal{f}},{\varepsilon}}s}=f^{{\varepsilon},s}_{a , b,*}(x)=0 $ ] in @xmath1577 . * step 2 : * let @xmath300 be the control pair of equation ( [ equation - compatiblility - f - boundary ] ) , then we have @xmath1578 since @xmath1579,\ldots,[p_l]$ ] belong to the prescribed system of generator for @xmath1512 and @xmath1580-[p'_1(0)],\dots,[p'_{l'}]-[p'_{l'}(0)]$ ] belong to the prescribed system of generator for @xmath1513 , we deduce from equation ( [ equ - diag ] ) and from the choice of @xmath130 that @xmath1581_{{\varepsilon},s}=0\]]for @xmath1570 and @xmath1582_{{\varepsilon},s}=0\]]for @xmath1572 . according to lemma [ lemma - bound1 ] and in view of the definition of @xmath1583 , then for a control pair @xmath1101 depending only on @xmath619 , there exists for @xmath1584 and up to replacing @xmath1550 by some @xmath1585 for some integer @xmath1586 two @xmath1587-@xmath1588-unitaries @xmath1589 and @xmath1590 respectively in @xmath1591 and @xmath1592 , with @xmath1593 and @xmath1594 respectively in @xmath1595 and @xmath1596 and such that @xmath1597 since we also have @xmath1598_{{\varepsilon},s}=0,\ ] ] according to proposition [ proposition - decom - unitaries - adjoint ] and up to rescaling @xmath1101 , we can assume that there exists two @xmath1587-@xmath1588-unitaries @xmath1599 and @xmath1600 respectively in @xmath1591 and @xmath1592 with @xmath1601 and @xmath1602 respectively in @xmath1595 and @xmath1596 and such that @xmath1603 and @xmath1604 is homotopic to @xmath1605 as an @xmath1587-@xmath1588-unitary in @xmath1606 for @xmath848 . in the same way , there exists two @xmath1587-@xmath1588-unitaries @xmath1607 and @xmath1608 in @xmath1609 and two @xmath1587-@xmath1588-unitaries @xmath1610 and @xmath1611 in @xmath1612 for some integer @xmath1613 such that @xmath1614 @xmath1615 and @xmath1616 is homotopic to @xmath1617 as an @xmath1587-@xmath1588-unitary in @xmath1618 for @xmath848 and @xmath1619 for @xmath1572 . in particular , @xmath1620 and @xmath1621 for @xmath848 and @xmath1570 are homotopic @xmath1587-@xmath1588-unitaries in @xmath1622 and @xmath1623 and @xmath1624 for @xmath848 and @xmath1572 are homotopic @xmath1587-@xmath1588-unitaries in @xmath1625 . * step 3 : * according to lemmas [ lemma - almost - closed ] and [ cor - example - homotopy ] and up to replacing @xmath22 by @xmath1626 , we have that @xmath1627_{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}{{\otimes}}[p_i])\\ & + \jmath^{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}_{{\delta}_2,*}{{\otimes}}id_{k_0(b)}(\sum_{i=1}^l[v_i^{(2)}]_{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}{{\otimes}}[p_i])\\ & + \jmath^{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}_{s{\delta}_1,*}{{\otimes}}id_{k_0(sb)}(\sum_{i=1}^{l'}[w_i^{(1)}]_{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}{{\otimes}}([p'_i]-[p'_i(0)]))\\ & + \jmath^{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}_{s{\delta}_2,*}{{\otimes}}id_{k_0(sb)}(\sum_{i=1}^{l'}[w_i^{(2)}]_{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}{{\otimes}}([p'_i]-[p'_i(0 ) ] ) ) \end{split}\ ] ] let us set for @xmath848 @xmath1628_{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}{{\otimes}}[p_i]\right)\\ & + \jmath^{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}_{s{\delta}_1,*}{{\otimes}}id_{k_0(sb)}\left(\sum_{i=1}^{l'}[w_i^{(j)}]_{\lambda\lambda_0{{\alpha}}{\varepsilon},2k_{{\lambda}_0{{\alpha}}{\varepsilon } } r'}{{\otimes}}([p'_i]-[p'_i(0)]\right ) \end{split}\ ] ] and @xmath1629 then @xmath1630 and @xmath1631 satisfy the required condition for some suitable control pair . _ end of the proof of the @xmath1507-statement . _ we divide the end of the proof of the @xmath1632-statement in the four following steps . * step 1 : * let @xmath18 be an element in @xmath1633 such that @xmath1634 . let @xmath130 be a positive number such that @xmath1635 . with notations of proposition [ proposition - technic ] , applying proposition [ proposition - decom - unitaries ] to @xmath1534 , up to rescaling @xmath261 and to replacing @xmath1636 and @xmath1631 respectively by @xmath1637 and @xmath1638 for some integer @xmath1639 , there exist for any positive number @xmath40 in @xmath1640 two @xmath766-@xmath1641-unitaries @xmath1642 and @xmath1643 in some @xmath1535 such that * @xmath1644 is an element in the matrix algebra @xmath1645 for @xmath650 ; * for @xmath650 , there exists a homotopy @xmath1646}$ ] of @xmath1238-@xmath1641-unitaries between @xmath160 and @xmath1647 such that @xmath1648 for all @xmath179 in @xmath180 $ ] . * @xmath1649 . up to replacing @xmath22 by @xmath1650 , we can assume indeed that @xmath1651 . if we apply the @xmath1652 property to @xmath1653 and @xmath1654 , we get that there exists @xmath1655 in @xmath1656 such that * @xmath1657 ; * @xmath1658 ; * @xmath1659 lies in @xmath1660 ; in particular , in view of lemma [ lemma - almost - closed ] , @xmath1655 is a @xmath1661-@xmath1662-unitary in @xmath1663 homotopic to @xmath1664 ( resp . to @xmath1665 ) as a @xmath1661-@xmath1662-unitary in @xmath1666 ( resp . in @xmath1667 ) . * step 2 : * let us define @xmath1668 if we set @xmath1669 , then * @xmath1670 is a @xmath1671-@xmath1672-unitary in @xmath1673 ; * @xmath1674 is in @xmath1675 ; * @xmath1676_{{\lambda}'{\varepsilon } , 4h_{\varepsilon}r'}$ ] lies in @xmath1677 ; then we have @xmath1678_{{\lambda}'{\varepsilon } , 4h_{\varepsilon}r'})=[{w_1'}^*w^{(1)}_x]_{{\lambda}'{\varepsilon } , 4h_{\varepsilon}r'}=[w^{(1)}_x]_{{\lambda}'{\varepsilon } , 4h_{\varepsilon}r'},\ ] ] where the the second equality holds because @xmath1642 is connected to @xmath160 as a @xmath1679-@xmath1672-unitary of @xmath1680 . in the same way , we have @xmath1681_{{\lambda}'{\varepsilon } , 4h_{\varepsilon}r'})=-[w^{(2)}_x]_{{\lambda}'{\varepsilon } , 4h_{\varepsilon}r'}.\ ] ] * step 3 : * let @xmath1317 be a positive integer with @xmath1682 such that for any @xmath365 in @xmath1683 , there exists @xmath19 in @xmath1684 such that @xmath1685 then there exists @xmath1686 in @xmath1687 such that @xmath1688_{{\lambda}_0{\lambda}'{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}_0 { \lambda}'{\varepsilon } } r''}= f^{{\lambda}_0{\lambda}'{\varepsilon},r''}_{a_{{\delta}_1}\cap a_{{\delta}_2},b,*}(x').\ ] ] hence we have @xmath1689_{{\lambda}_0{\lambda}'{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}_0 { \lambda}'{\varepsilon } } r''})\\ \nonumber&= & [ w^{(1)}_x]_{{\lambda}_0{\lambda}'{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}_0 { \lambda}'{\varepsilon } } r''}\\ & = & f^{{\lambda}_0{\lambda}'{\varepsilon},r''}_{a_{{\delta}_1},b,*}\circ \iota_{{\mathcal{o}},*}^{{\lambda}_0{\lambda}'{\varepsilon},r''}(x^{(1)})\end{aligned}\ ] ] where the first equality holds by naturality of @xmath1440 , the second equality holds in view of equation ( [ equ - vx1 ] ) and the last equality is a consequence of proposition [ proposition - technic ] . in the same way , using equation ( [ equ - vx2 ] ) , we get that @xmath1690 * step 4 : * let @xmath606 be a positive number , with @xmath1691 such that @xmath1692 for all @xmath19 in @xmath1693 such that @xmath1694 and @xmath848 . in particular , from equations ( [ equ1-end - qi ] ) and ( [ equ2-end - qi ] ) , we deduce @xmath1695 and @xmath1696 since @xmath1697 , this establishes the @xmath1507-statement . let @xmath864 be an @xmath6-controlled nuclear mayer - vietoris pair with coercitivity @xmath619 for @xmath4 such that for any @xmath2-algebra @xmath232 with @xmath1325 free abelian , then @xmath1441 and @xmath1442 are quantitative isomorphisms with rescaling @xmath301 . let us check that @xmath1443 satisfies the @xmath1698-condition of definition [ def - quantitative - morphism ] . as for the @xmath1507-condition , we can assume without loss of generality that @xmath4 and @xmath232 are unital and that there exists a system of generators of @xmath1512 given by classes of projections . let us fix such a system of generator for @xmath1512 and fix also a system of generator for @xmath1699 . we also only need to consider the odd case , i.e show that the control morphism @xmath1514 sastifies the required conditions . let @xmath40 be a positive number with @xmath40 in @xmath1700 and let @xmath1701 be a positive number with @xmath1702 such that for any @xmath19 in in @xmath1703 , there exists an element @xmath18 in @xmath1704 such that @xmath1705 let @xmath1706 be a positive number with @xmath1707 such that @xmath1708 for all @xmath18 in @xmath1709 such that @xmath1710 . [ proposition - technic - qs ] there exists a control @xmath261 depending only on @xmath619 and @xmath301 with @xmath1711 such that for any positive numbers @xmath40 and @xmath130 with @xmath40 in @xmath1640 and @xmath1712 the following is satisfied : for any element @xmath19 in @xmath1713 , there exists * an element @xmath1714 in @xmath1715 ; * a positive integer @xmath1716 and two @xmath1679-@xmath1717-unitaries @xmath1718 and @xmath1719 in @xmath1720 ; * for @xmath848 a @xmath1238-@xmath1641-unitary @xmath1721 in @xmath1722 with @xmath1723 in @xmath1724 ; * a @xmath1238-@xmath1641-projection @xmath1725 in @xmath1726 such that * @xmath1727_{{\lambda}{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}{\varepsilon}}h_{\varepsilon}r'}$ ] is in @xmath1728 and @xmath1729_{{\lambda}{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}{\varepsilon}}h_{\varepsilon}r'}$ ] ; * @xmath1730_{{\lambda}{\varepsilon},h_{\varepsilon}r'}=\iota^{{{\alpha}}_c{\varepsilon},{\lambda}{\varepsilon},k_{c,{\varepsilon}}r , h'_{\varepsilon}r'}_{{\mathcal{o}}',*}\circ\partial^{{\varepsilon},r}_{c({{\mathbb t}}_2,{\delta}_1{{\otimes}}b),c({{\mathbb t}}_2,{\delta}_2{{\otimes}}b)}(y)$ ] . * @xmath1731 ; * @xmath1732 the proof of this proposition is also quite long so is divided into four steps . * step 1 : * according to equation ( [ equation - boundary - eight ] ) , we see that @xmath1733 belongs to @xmath1734 . set @xmath1735 . then there exists an element @xmath18 in @xmath1736 such that @xmath1737 there exist two integers @xmath87 and @xmath102 and * for @xmath1549 * * an @xmath1738-@xmath606-projection @xmath1739 in some @xmath1740 and an integer @xmath1741 . * * a projection @xmath1552 in some @xmath1553 with @xmath1554 $ ] in the prescribed system of generators for @xmath1512 ; * for @xmath1555 * * an @xmath1738-@xmath606-projection @xmath1742 in some @xmath1743 and an integer @xmath1744 ; * * a projection @xmath1558 in some @xmath1559 such that @xmath1560-[p'_i(0)]$ ] is in the prescribed system of generators for @xmath1561 ; such that @xmath1745_{{\lambda}_0{{\alpha}}_c{\varepsilon},r}{{\otimes}}[p_i]+\sum_{i=1}^{l'}[q'_i , k'_i]_{{\lambda}_0{{\alpha}}_c{\varepsilon},r}{{\otimes}}([p'_i]-[p'_i(0)]).\ ] ] by morita equivalence , up to replacing @xmath232 by @xmath1566 with @xmath1568 , we can assume that @xmath1746 for @xmath1570 and @xmath1747 for @xmath1572 . using lemma [ lemma - almost - canonical - form ] , we can moreover assume without loss of generality that @xmath1748 is in @xmath1749 for @xmath1570 and that @xmath1750 for @xmath1572 . set @xmath1751 in view of equations ( [ equ - f - even ] ) and ( [ equ - f - even - suspension ] ) , @xmath1752_{{\lambda}_0{{\alpha}}_c{{\alpha}}_{\mathcal{f}}{\varepsilon},k_{{\mathcal{f}},{\lambda}_0{{\alpha}}_c}r},\ ] ] with @xmath1753 . by naturality of @xmath1440 , we obtain @xmath1754 in @xmath1755 and @xmath1756 in @xmath1757 . let @xmath130 be a positive number with @xmath1712 . since @xmath1554 $ ] for @xmath1570 and @xmath1560-[p'_i(0)]$ ] for @xmath1572 are respectively in the prescribed system of generator of @xmath1512 and @xmath1699 and by definition of @xmath1706 , we deduce that * @xmath1758_{{\lambda}^2_0{{\alpha}}_c{\varepsilon},r'})=0 $ ] in @xmath1759 for @xmath1760 ; * @xmath1761_{{\lambda}^2_0{{\alpha}}_c{\varepsilon},r'})=0 $ ] in @xmath1762 for @xmath1763 ; * @xmath1764_{{\lambda}^2_0{{\alpha}}_c{\varepsilon},r'})=0 $ ] in @xmath1765 for @xmath1760 ; * @xmath1766_{{\lambda}^2_0{{\alpha}}_c{\varepsilon},r'})=0 $ ] in @xmath1767 for @xmath1763 . let @xmath261 be the control pair of lemma [ lemma - technic2 ] . then for @xmath1549 and up to stabilisation , we can assume that @xmath1768 and that there exists @xmath1589 and @xmath1590 two @xmath1769-@xmath1770 unitaries in @xmath1771 and @xmath1550 and @xmath1556 two @xmath1769-@xmath1770 unitaries in @xmath1551 such that * @xmath1772 is an element in @xmath1773 for @xmath848 ; * @xmath1774 and @xmath1775 * for @xmath848 , then @xmath1776 is connected to @xmath1777 by a homotopy of @xmath1769-@xmath1778-unitaries @xmath1779}$ ] in @xmath1551 such that @xmath1780 is in @xmath1773 for all @xmath179 in @xmath180 $ ] . * @xmath1781 . in the same way , for @xmath1555 and up to stabilisation , we can assume that @xmath1782 and that there exists @xmath1783 and @xmath1784 two @xmath1769-@xmath1770 unitaries in @xmath1557 and @xmath1785 and @xmath1786 two @xmath1769-@xmath1787 unitaries in @xmath1557 such that * @xmath1788 is an element in @xmath1789 for @xmath848 ; * @xmath1790 and @xmath1791 * for @xmath848 , then @xmath1792 is connected to @xmath1793 by a homotopy of @xmath1769-@xmath1770-unitaries @xmath1794}$ ] in @xmath1557 such that @xmath1795 is in @xmath1789 for all @xmath179 in @xmath180 $ ] . * @xmath1796 ; * @xmath1797 . * step 3 : * let us set @xmath1798 and for any @xmath40 in @xmath1799 set @xmath1800 . consider the element of @xmath1801 @xmath1802_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}{{\otimes}}[p_1]+\ldots+[u_l]_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}{{\otimes}}[p_l]+[u''_1]_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}{{\otimes}}([p'_1]-[p'_1(0 ) ] ) + \ldots+[u''_{l'}]_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}{{\otimes}}([p'_{l'}]- [ p'_{l'}(0)]).\ ] ] if we set @xmath1803 then @xmath1718 is a @xmath1679-@xmath1717-unitary in @xmath1722 . using lemmas [ lemma - almost - closed ] and [ cor - example - homotopy ] and up to replacing @xmath1804 by @xmath1805 , then for @xmath1549 ( resp . for @xmath1572 ) @xmath1806 is homotopic to @xmath1807 ( resp . @xmath1808 is homotopic to @xmath1809 ) as a @xmath1679-@xmath1717-unitary in @xmath1551 ( resp . in @xmath1557 ) . hence we deduce that @xmath1727_{{{\alpha}}_{\mathcal{f}}{\lambda}'{\varepsilon},k_{{\mathcal{f}},{\lambda}'{\varepsilon}}h'_{\varepsilon}r'}$ ] belongs to @xmath1810 and in view of equations ( [ equ - f ] ) and ( [ equ - f - suspension ] ) , we see that @xmath1811_{{{\alpha}}_{\mathcal{f}}{\lambda}'{\varepsilon},k_{{\mathcal{f}},{\lambda}'{\varepsilon}}h'_{\varepsilon}r'}=f^{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}_{a , b,*}(z_y)\ ] ] in @xmath1810 . in the same way , if we set @xmath1812 then @xmath1719 is also an @xmath1679-@xmath1717-unitary in @xmath1722 . * step 4 : * for @xmath1570 and @xmath848 , let @xmath1604 be the matrix in @xmath1551 obtained from @xmath1813 by flipping the @xmath1741 first and the @xmath1741-last coordinates , and define similarly @xmath1616 in @xmath1814 for @xmath1572 and @xmath848 . up to replacing @xmath1815 by @xmath1816 , we have that * @xmath1817 and @xmath1818 are @xmath1679-@xmath1819 unitaries in @xmath1551 ; * @xmath1820 and @xmath1821 * @xmath1822 . where @xmath1823 is obtained from @xmath1739 by flipping the @xmath1741 first and the @xmath1741 last coordinates . similarly , for @xmath1555 , we have * @xmath1608 and @xmath1611 are two @xmath1679-@xmath1819 unitaries in @xmath1557 ; * @xmath1824 is an element in @xmath1825 for @xmath848 ; * @xmath1826 and @xmath1827 * @xmath1828 . where @xmath1829 is obtained from @xmath1742 by flipping the @xmath1744 first and the @xmath1744 last coordinates . then we have @xmath1830 and @xmath1831 with @xmath1832 for @xmath1833 . similarly , we have @xmath1834 and @xmath1835 for @xmath1836 with @xmath1837 from this we deduce that there exist @xmath1838 and @xmath1839 two @xmath1815-@xmath1717-unitaries in @xmath1722 such that * @xmath1840 is in @xmath1841 for @xmath1842 ; * @xmath1843 ; * @xmath1844 , where @xmath1845 clearly @xmath1725 is @xmath1679-@xmath1717-projection in @xmath1722 such that @xmath1846_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}=[{q'_y},n_y]_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}=\iota^{{\varepsilon},{\lambda}'{\varepsilon},r , h'_{\varepsilon}r'}_{{\mathcal{o}}',*}\circ\partial^{{\varepsilon},r}_{c({{\mathbb t}}_2,{\delta}_1{{\otimes}}b),c({{\mathbb t}}_2,{\delta}_2{{\otimes}}b)}(y)\ ] ] and hence @xmath1847 and @xmath1839 satisfy the required conditions for some suitable control pair . _ end of the proof of the @xmath1698-statement . _ let @xmath261 be a control pair as in proposition [ proposition - technic - qs ] , let @xmath40 be a positive number in @xmath1640 , let @xmath19 be an element in @xmath1848 and let @xmath1849 and @xmath1839 as in the proposition . let @xmath42 be an @xmath40-@xmath6 unitary in some @xmath1535 , let @xmath1153 and @xmath1154 be @xmath1850-@xmath1851-unitaries in some @xmath1852 and let @xmath53 be an @xmath1850-@xmath1853-projection in @xmath1852 such that * @xmath1854 is in @xmath1855 for @xmath650 ; * @xmath1856 ; * @xmath1057 is in @xmath1857 * @xmath1858 ; * @xmath1859 ; * @xmath1860_{{\varepsilon},r}$ ] ; * @xmath1861_{{{\alpha}}_c{\varepsilon},k_{c,{\varepsilon } } r}$ ] . then applying lemma [ lemma - bound1 ] to @xmath1862 and to the matrices respectively obtained from @xmath1863 and @xmath1864 by swapping the order of coordinates @xmath1865 and @xmath1866 , we see that for a controlled pair @xmath1804 depending only on @xmath301 and @xmath619 , and if @xmath40 is in @xmath1799 , there exist @xmath1867 and @xmath1868 some @xmath1679-@xmath1717-unitary in some @xmath1869 with @xmath1870 in @xmath1871 and @xmath1872 in @xmath1873 such that @xmath1874_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}+[u_2]_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}=[w_{y}]_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}-\iota_{{\mathcal{o}}',*}^{{\varepsilon},{\lambda}'{\varepsilon},r , h'_{\varepsilon}r'}(y)\ ] ] in @xmath1875 . up to replacing @xmath1876 for @xmath848 by @xmath1877 and @xmath1804 by @xmath1878 , we can assume that @xmath1879_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'}$ ] belongs to @xmath1880 . let @xmath1317 be a positive number with @xmath1881 such that for @xmath848 , any positive number @xmath40 in @xmath1882 and any @xmath365 in @xmath1883 , there exists an element @xmath18 in @xmath1884 such that @xmath1885let then @xmath1886 be for @xmath848 an element in @xmath1887 such that @xmath1888_{{\lambda}'{\varepsilon},h'_{\varepsilon}r'})=f^{{\lambda}_0{\lambda}'{\varepsilon},r''}_{a_{{\delta}_j},b,*}(z_{y}^{(j)}).\ ] ] let us set @xmath1889 in @xmath1890 . by naturality of @xmath1440 , we see then that @xmath1891 and hence the @xmath1698-condition is satisfied . we shall next discuss the connection between the baum - connes conjecture and the quantitative knneth formula . the connection between the usual baum - connes conjecture and the knneth formula was studied in @xcite . before proving theorem [ thm - bc - qkunneth ] , recall that article @xcite introduced an equivariant analogue of the map @xmath1892 for the topological @xmath0-theory of a locally compact group @xmath1893 ( i.e. , the left - hand side of the baum - connes assembly map ) . let @xmath4 be a @xmath1893-@xmath2-algebra and let @xmath232 be a @xmath2-algebra . the @xmath2-algebra @xmath232 can be viewed as a @xmath1893-@xmath2-algebra with the trivial action of @xmath1893 and we equip @xmath357 with the diagonal action . then the elements in @xmath1325 can be viewed as element of @xmath1894 . if @xmath668 is a @xmath1893-proper space , the map @xmath1895 is compatible with inclusion of @xmath1893-proper cocompact spaces and hence gives rise to a morphism @xmath1896 [ thm - bc - kunneth ] let @xmath393 be a discrete group and let @xmath4 be a @xmath393-@xmath2-algebra . assume that for every finite subgroup @xmath1300 of @xmath393 , the @xmath2-algebra @xmath1897 satisfies the knneth formula , then for any @xmath2-algebra @xmath232 such that @xmath1325 is a free abelian group and any positive number @xmath439 @xmath1898 is an isomorphism . ( compare with the proof of ( * ? ? ? * lemma 1.7 ) ) the action of @xmath393 on @xmath1899 is simplicial and up take a barycentric subdivision of @xmath1900 , we can assume that @xmath1900 is a locally finite and finite dimension typed simplicial complex , equipped with a simplicial and type preserving action of @xmath393 . let @xmath1901 be the skeleton decomposition of @xmath440 . then @xmath1902 is a simplicial complex of dimension @xmath1639 , locally finite and equipped with a proper , cocompact and type preserving simplicial action of @xmath393 . let us prove by induction on @xmath1639 that @xmath1903 is an isomorphism . the @xmath1133-skeletton @xmath1904 is a finite union of orbits and thus , for @xmath1905 , it is enought to prove that @xmath1906 is an isomorphism when @xmath1300 is a finite subgroup of @xmath393 . let us recall from @xcite that for every @xmath2-algebra @xmath232 equipped with an action of @xmath393 , there is a natural restriction isomorphism @xmath1907 moreover , we have the following commutative diagram @xmath1908 the bottom row being by assumption an isomorphism , the top row is then also an isomorphism . let us assume that we have proved that @xmath1909 is an isomorphism . then the short exact sequence @xmath1910 gives rise to an natural long exact sequence @xmath1911 and thus by naturality and since @xmath1325 is a free abelian group , we get a commutative diagram let @xmath1912 be the interior of the standard @xmath1639-simplex . since the action of @xmath393 is type preserving , then @xmath1913 is equivariantly homeomorphic to @xmath1914 , where @xmath1915 is the set of center of @xmath1639-simplices of @xmath1916 , and where @xmath393 acts trivially on @xmath1912 . this identification , together with bott periodicity , provides a commutative diagram @xmath1917 by the first step of induction , @xmath1918 is an isomorphism , and hence @xmath1919 is an isomorphism . using the induction hypothesis and the five lemma , we conclude that @xmath1920 is an isomorphism . [ lemma - kunneth ] there exits a positive number @xmath301 and a function @xmath1921 non decreasing in @xmath439 and non increasing in @xmath40 with @xmath1922 for all @xmath40 in @xmath1923 and @xmath1924 such that the following holds : for any finitely generated group @xmath393 , any @xmath394-@xmath2-algebra @xmath4 , any @xmath2-algebra @xmath232 and any positive numbers @xmath40 , @xmath6 and @xmath439 with @xmath206 and @xmath1925 , then we have @xmath1926 let @xmath365 be an element in @xmath1927 and let @xmath19 be an element in @xmath1325 . then @xmath1928_{\frac{{\varepsilon}}{\alpha_j},\frac{r}{k_{j,{{\varepsilon}}/{\alpha_j}}}}\right){{\otimes}}y\right)&\\ = \tau_{a{{\rtimes_{red}}},*}^{{\varepsilon},r}(y)\circ j_\gamma^{red,\frac{{\varepsilon}}{\alpha_j},\frac{r}{k_{j,{{\varepsilon}}/{\alpha_j}}}}&(z)\left([p_{{{\gamma}},d},0]_{\frac{{\varepsilon}}{\alpha_j},\frac{r}{k_{j,{{\varepsilon}}/{\alpha_j}}}}\right).\end{split}\ ] ] the result is then a consequence of remark [ remark - kasp - tens ] . + _ proof of theorem [ thm - bc - qkunneth ] . _ let @xmath200 and @xmath1929 as in lemma [ lemma - kunneth ] . let @xmath1930 be a positive number as in as in theorem [ thm - quant - surj ] , let @xmath40 and @xmath6 be positive numbers with @xmath1931 . let @xmath439 and @xmath606 be positive numbers with @xmath1932 such that @xmath1933 is satisfied for every @xmath393-@xmath2-algebra @xmath4 . let @xmath462 be a positive number such that @xmath1934 is satisfied for every @xmath393-@xmath2-algebra @xmath4 . * let @xmath19 be an element in @xmath1935 . since @xmath1936 and @xmath1937 is satisfied , there exists an element @xmath1938 in @xmath1939 such that @xmath1940 according to theorem [ thm - bc - kunneth ] there exists @xmath1941 in @xmath1942 such that @xmath1943 . now if we set @xmath1944 and @xmath1945 , then @xmath1946 is in @xmath1947 and from lemma [ lemma - kunneth ] we deduce that @xmath1948 * let @xmath18 be an element in @xmath1949 such that @xmath1950 in @xmath1951 , let @xmath1941 be an element in @xmath1942 such that @xmath1952 and let us set @xmath1953 according to lemma [ lemma - kunneth ] , we have that @xmath1954 in @xmath1955 and hence since @xmath1956 is satisfied , we have @xmath1957 in @xmath1958 . according to theorem [ thm - bc - kunneth ] and since @xmath1959 is compatible with inclusion @xmath1960 we deduce that @xmath1961 in @xmath1962 . set @xmath1963 and pick any positive number @xmath130 such that @xmath1964 and @xmath1965 . then we have @xmath1966 in this section , we introduce the concept of finite asymptotic nuclear decomposition for filtered @xmath2-algebras and and we prove for this class of @xmath2-algebras the quantitative knneth formula . we deduce from this that uniform roe algebras of discrete metric spaces with bounded geometry and finite asymptotic dimension satisfy the knneth formula . 1 . @xmath1967 contains @xmath185 ; 2 . @xmath1967 is closed under countable inductive limits ; 3 . @xmath1967 is stable under extension , i.e for any extension of @xmath2-algebras @xmath1968 if any two of the @xmath2-algebras are in @xmath1967 then so is the third ; 4 . @xmath1967 is closed under @xmath354-equivalence . a filtered @xmath2-algebra @xmath4 with filtration @xmath7 is called locally bootstrap if for all positive number @xmath143 there exists a positive number @xmath6 with @xmath1969 and a sub-@xmath2-algebra @xmath1970 of @xmath4 such that * let us prove first the @xmath1972-statement . let @xmath40 and @xmath143 be positive numbers with @xmath1973 . let then @xmath6 be a positive number with @xmath1974 and let @xmath1970 be a @xmath2-algebra such that @xmath1970 belongs to the bootstrap class and @xmath1975 . then @xmath1970 is filtered by @xmath1976 and the filtration is indeed finite , i.e. @xmath1977 for any positive number @xmath242 with @xmath1978 . let us consider the commutative diagram @xmath1979where the left bottom and left top maps are induced by the inclusion @xmath1980 for any @xmath1981 . let @xmath18 be an element in @xmath1982 such that @xmath1983 and let then @xmath19 in @xmath1984 be the image of @xmath18 under the compositions of the top row . then @xmath1985 and hence @xmath1986 since @xmath1970 is in the bootstrap class , then @xmath1987 is an isomorphism and hence @xmath1988 in @xmath1989 . since @xmath1325 is free abelian and according to proposition [ proposition - approximation - k - th ] , there exists a positive number @xmath130 , with @xmath1990 such that @xmath1991 in @xmath1992 . but since @xmath1970 has propagation less than @xmath6 , then @xmath1993 in @xmath1994 . hence composing with the map @xmath1995 induced by the inclusion @xmath1996 , we get then that + @xmath1997 * let us prove now the @xmath1998-statement . let @xmath143 and @xmath40 be positive numbers with @xmath1999 and let @xmath1970 be a @xmath2-algebra such that @xmath1970 belongs to the bootstrap class and @xmath2000 . let @xmath365 be an element in some @xmath2001 and let @xmath371 in @xmath2002 be the image of @xmath365 under the map @xmath2003 induced by the inclusion @xmath2004 . since @xmath1970 is in the bootstrap class , there exists @xmath19 in @xmath1989 such that @xmath2005 in @xmath2006 . since any element of @xmath1970 has propagation less than @xmath6 , there exists an element @xmath18 in @xmath2007 such that @xmath2008 in @xmath2009 . since @xmath2010we get that @xmath2011 and @xmath2012 have same image under the map @xmath2013 hence , according to proposition [ proposition - approximation - k - th ] , there exists a positive number @xmath130 , with @xmath2014 , such that @xmath2015 but since @xmath2016 for all @xmath2017 we get that @xmath2018 composing with the map @xmath2019 induced by the inclusion @xmath1996 , we get then that @xmath2020 where @xmath1686 is the image of @xmath2021 under the composition @xmath2022 induced by the inclusion @xmath1996 . a family of filtered @xmath2-algebras @xmath2023 is uniformly locally bootstrap if for all integer @xmath311 and for all positive number @xmath143 , there exist a positive number @xmath6 with @xmath1969 and a sub-@xmath2-algebra @xmath2024 of @xmath2025 such that for all integer @xmath311 , [ prop - uniformly - locally - bootstrap]there exists a positive number @xmath301 with @xmath302 such that any uniformly locally bootstrap family @xmath2023 of filtered @xmath2-algebras and any @xmath2-algebra @xmath232 with @xmath1325-free abelian then @xmath2028 is a uniform family of quantitative isomorphisms with rescalling @xmath301 let us define @xmath2029 as the class of uniformly locally bootstrap families of @xmath2-algebras . then we define by induction @xmath2030 as the class of family @xmath2031 for which there exists a positive number @xmath619 such that for every positive number @xmath6 , the following is satisfied : there exists a family @xmath2032 in @xmath2033 and for any @xmath2-algebra @xmath4 in @xmath2031 an @xmath6-controlled nuclear mayer - vietoris pair @xmath864 with coercitivity @xmath619 for @xmath4 with @xmath1303 and @xmath832 in @xmath2032 . define then @xmath2034 as the class of families @xmath2035 such that @xmath2035 is in @xmath2030 for some integer @xmath5 . theorem [ thm - mv - kunneth ] obviously admits a uniform version for families and hence , together with proposition [ prop - locally - bootstrap ] , we obtain the following result . [ prop - fnc ] let @xmath2035 be a family in @xmath2034 . then there exists a positive number @xmath2036 with @xmath2037 such that for any @xmath2-algebra @xmath232 with @xmath1325 free abelian , then @xmath2038 is a uniform family of quantitative isomorphisms with rescalling @xmath2036 ( indeed @xmath2035 only depends on @xmath5 such that @xmath2035 lies in @xmath2030 . typical examples of family of filtered @xmath2-algebra in @xmath2034 are provided by spaces with asymptotic dimension . recall that for a metric space @xmath668 and a positive number @xmath6 , a cover @xmath2042 has @xmath6-multiplicity @xmath5 if any ball of radius @xmath6 in @xmath668 intersects at most @xmath5 elements in @xmath2042 . let @xmath2043 be a proper discrete metric space . then @xmath2043 has * finite asymptotic dimension * if there exists an integer @xmath998 such that for any positive number @xmath6 , there exists a uniformly bounded cover @xmath2042 with finite @xmath6-multiplicity @xmath2044 . the smallest integer that satisfies this condition is called the * asymptotic dimension * of @xmath495 . 1 . @xmath495 has asymptotic dimension @xmath998 ; 2 . for every positive number @xmath6 there exist @xmath2044 subsets @xmath2045 of @xmath495 such that * @xmath2046 ; * for @xmath2047 , then @xmath2048 is the @xmath6-disjoint union of a family @xmath2049 of subsets of @xmath2048 with uniformly bounded diameter , i.e @xmath2050 if @xmath2051 and there exists a positive number @xmath615 such @xmath2052 for all integer @xmath86 . let @xmath495 be a proper metric space with asymptotic dimension @xmath998 , then there exists a sequence of positive numbers @xmath2053 and for any integer @xmath86 a cover @xmath2054 of @xmath495 such that * @xmath2055 for every integer @xmath86 ; * @xmath2056 has diameter less than @xmath2057 for every integers @xmath311 and @xmath86 ; * for any integer @xmath86 , the @xmath2057-multiplicity of @xmath2058 is @xmath2044 . [ lemma - afd ] let @xmath998 be an integer , let @xmath2053 be a sequence of positive numbers such that @xmath2060 for every integer @xmath86 . let @xmath2061 be a family of proper metric spaces with asymptotic dimension @xmath998 and @xmath998-growth @xmath2053 . then for any family @xmath2062 in the bootstrap category , the family @xmath2063 belongs to @xmath2034 . let us equip @xmath2064 with a distance @xmath2065 such that the inclusion @xmath2066 are isometric for all integer @xmath311 and @xmath2067 for all integers @xmath311 and @xmath1639 with @xmath801 . then @xmath495 has asymptotic dimension @xmath998 and hence according to @xcite , the metric space @xmath495 embeds uniformly in a product of trees @xmath2068 . let @xmath439 be the metric on @xmath2069 and @xmath2070 the distance on @xmath2071 when @xmath311 runs through integers . then there exists two non - decreasing functions @xmath2072 and for every integer @xmath311 a map @xmath2073 such that if @xmath2076 , then @xmath668 is a tree and then the result holds in view of examples [ example - roe - algebras ] , [ example - roe ] and [ example - tree ] . a straightforward induction shows that if @xmath495 embeds uniformly in a product of @xmath5 trees , then @xmath2063 is in @xmath2034 . let @xmath2077 be a family of filtered @xmath2-algebras . then the uniform products of @xmath2077 , denoted by @xmath2078 is the closure of @xmath2079 in @xmath2080 equipped with the supremium norm . the uniform product @xmath2078 is then obviously a filtered @xmath2-algebra . it was proved in ( * ? ? ? * lemma 1.11 ) that the quantitative @xmath0-theory of a uniform product of a stable filtered @xmath2-algebra is computable in term of the quantitative @xmath0-theory of the algebras of the family . a @xmath2-algebra is said to be of finite asymptotic nuclear @xmath2081-decomposition if there exists a positive number @xmath619 and an integer @xmath5 such that for any positive number @xmath6 , there exists a @xmath6-controlled mayer - vietoris pair @xmath864 that satisfies the following . there exist three families of filtered @xmath2-algebras @xmath2082 and @xmath2083 in @xmath2034 such that @xmath1303 and @xmath832 are respectively isomorphic as filtered @xmath2-algebras to @xmath2084 and @xmath2085 . if @xmath495 is a proper metric set of bounded geometry and with finite asymptotic dimension . then the uniform roe algebra @xmath2086 has asymptotic finite nuclear @xmath2081-decomposition and the roe algebra @xmath797 has stably asymptotic finite nuclear @xmath2081-decomposition . let us prove de result for @xmath2086 , the proof for @xmath797 being similar . let us fix @xmath2087 in @xmath495 and let @xmath6 be a positive number . let us fix @xmath143 and @xmath606 two positive numbers such that @xmath2088 . set for @xmath86 integer @xmath2089 and @xmath2090 then @xmath842 and @xmath2048 is for @xmath650 the @xmath606-disjoint union of the family @xmath2049 . let @xmath2091 be for @xmath650 the set of element in @xmath2086 with support in @xmath2092 since @xmath495 has bounded geometry , then with notations of example [ example - roe - algebras ] , there exists a controlled mayer - vietoris @xmath864 of order @xmath6 and coercitivity @xmath13 such that @xmath2093 for @xmath650 and @xmath2094 the result is now a consequence of lemma [ lemma - afd ] . if @xmath495 is a discrete proper metric set of bounded geometry and with finite asymptotic dimension . then the uniform roe algebra @xmath2086 satisfies the quantitative knneth formula for some rescaling @xmath22 .

in this paper , we apply quantitative operator @xmath0-theory to develop an algorithm for computing @xmath0-theory for the class of filtered @xmath1-algebras with asymptotic finite nuclear decomposition . as a consequence , we prove the knneth formula for @xmath1-algebras in this class . our main technical tool is a quantitative mayer - vietoris sequence for @xmath0-theory of filtered @xmath1-algebras . _ keywords : quantitative operator @xmath0-theory , knneth formula , filtered @xmath1-algebras _ _ 2000 mathematics subject classification : 19k35,46l80,58j22 _

the complexity of the hadronic structure is one of the actual and topical fields of physics concerning the microscopic scale . different , complementary as well as competing experimental approaches are used to attack the challenge of the hadronic spectroscopy . a detailed understanding of the strong interaction dynamics on the symmetries of qcd is a fundamental aim of hadron physics . at high energies the qcd serves as a reliable theory whereas at low energies the non perturbative qcd phenomena must be investigated . here calculations are performed in an effective field theory of hadrons represented by the chiral effective lagrangian , a theory of goldstone bosons coupled to the octet of baryons and vector mesons . the qcd lagrange function is written as : @xmath1 in principle , @xmath2 describes the column vector of the quarks @xmath3 with each of them represented by a spinor , identical to the dirac equation for spin@xmath4 electrons . at energies in the few gev range only the first three quarks are of any relevance , @xmath5 represents the diagonal of the matrix for the quark current masses . data of high quality and precision on hadronic processes at low and intermediate momenta are necessary in order to verify the systematic low energy expansion of the chiral perturbation theory ( @xmath6 ) , which has already enforced an important insight into the structure and dynamics of nucleons and mesons @xcite . it is known that the current masses of the lightest three quarks are significantly smaller than the typical hadronic energy scale , represented by the proton mass of @xmath7 . this is the reason why in first order the quark masses of the light quarks are neglected in theoretical considerations and the spin@xmath4 fields can be separated into two independent left and right handed parts , reflected by the chiral symmetry . the qcd lagrangian separates into two identical images : @xmath8 = l_{qcd}\;[\bar q_{left } , q_{left } ] + l_{qcd}\;[\bar q_{right } , q_{right}].\ ] ] left handed quarks do not communicate with right handed quarks and _ vice versa_. there is no lorentz transformation which can change the handedness of a mass less quark or particle . this symmetry of the lagrange function is not the symmetry of the spectrum of the particles since otherwise each hadron should have a partner of equal mass but inverted parity . consequently there would be a second proton with negative parity which does not exist . there are no parity doublets in the spectrum of the strong interacting particles and therefore the symmetry is spontaneously broken . due to this breaking the mass less so called goldstone bosons , the pseudoscalar @xmath9 , @xmath10 and @xmath11 mesons appear . the restriction of the interaction of these particles among each other and with other hadrons can be used to analyze consequences of the chiral symmetry and its breaking in the framework of an effective field theory . in reality the lightest hadrons the goldstone bosons are by no means mass less , which can be understood by an explicit breaking of the chiral symmetry including a term for the quark masses . for each massive fermion there exists a lorentz transformation which transforms a left handed to a right handed field . the effect of this explicit breaking now can be treated perturbatively since the masses of the light quarks are much smaller than the typical hadronic mass . predicting the low energy properties of nuclear and particle physics , @xmath6 plays an important role in understanding hadron physics in the non perturbative regime . therefore it is very important to know the properties of the mesons , their structure and interaction in the hadronic environment . the physics program at the medium energy hadron accelerators was and is focusing on studies of the production and the decay of light mesons and baryon resonances and the conservation or violation of symmetries . following this sequence we will present and discuss production of mesons and meson pairs at threshold with the questions depicted by the interaction view of figure [ meson_exchange ] @xcite where in the nucleon nucleon ( @xmath12 ) scattering a meson @xmath13 is created in a one boson exchange model . for the particular case the questions have to be answered : how is the distortion of the incident @xmath12 waves ( isi ) included , which mesons contribute to the exchange process , is there an intermediate baryon resonance , how significant are rescattering contributions of the exchange mesons and what is a reasonable treatment of the @xmath12 and @xmath14 final state interactions ( fsi ) ? especially for the @xmath12interaction the fsi is crucial because of the nearby poles in the s wave amplitudes corresponding to the deuteron bound state in the @xmath15 channel or the @xmath16 virtual state @xcite . these poles and the phase space factors tend to determine much of the energy dependence of the total cross section for meson production . furthermore , in any region where these poles dominate , it is possible to link quantitatively meson production in cases where the two nucleons emerge separately or as a bound deuteron state . even at threshold the reaction mechanism of the basic process for the interrelation between real pions and virtual exchange meson currents as for instance in the pion production ( @xmath17 ) is still not fully understood . though first data suggested that s wave pions were produced in a heavy meson exchange process , later measurements resulted in an interpretation of an interference among transition amplitudes as ss , ps , pp , sd and ds , where the capital letter indicates the @xmath12 final state wave and the small letter the angular momentum between the two nucleons and the meson produced . in addition , it has been concluded that for higher partial waves @xmath9 exchange rather than a heavier meson exchange is more significant . especially the pion nucleon interaction went through several stages of increasing sensitivity for tests of nuclear theory . as indicated above more than a decade after the initial work @xcite only in the sixties explicit calculations on cross sections of the @xmath18 transitions were possible @xcite which , however , failed the total cross section @xcite by a factor of five . after further theoretical developments , see e.g. reference @xcite , only recently detailed calculations were provided within the jlich meson exchange model @xcite including transitions in the threshold region beyond @xmath19 and predicting analyzing powers and spin correlation coefficients making use of i ) the basic diagrams , ii ) final state interactions , iii ) off shell effects , iv ) the exchange of heavier mesons and v ) influences of intermediate excitation of the @xmath20 resonance . on the other hand , the jlich model does not account for quark degrees of freedom i.e. calculations within the framework of the @xmath6 which , in fact , predict cross sections too small @xcite compared to experiments.the heavier the meson produced in nucleon nucleon scattering the larger the momentum transfers and thus in such processes the short range parts of the production operators are tested . double meson production processes with both mesons being either identical ( e.g. @xmath21 or @xmath22 ) or different ( e.g. @xmath23 ) are in principle similar to the single ones , however , the possibly associated baryon resonances as intermediate states differ significantly . for instance , the two pion production very likely is dominated by either the @xmath24 @xmath25 resonance via an effective @xmath26 exchange giving information of a scalar meson excitation of the @xmath24 @xmath25 resonance or the simultaneously excited @xmath27 @xmath28 resonances . due to selection rules here the choice of definite quantum numbers in the initial and final state might help to sort out the reaction mechanism . in any case , a large amount of the knowledge about the @xmath9@xmath9 interaction has been obtained by the @xmath29 reaction , whereas data employing proton beams just start to come . as long as only s wave processes are involved , which is the privilege of threshold production studies , scattering length and effective range approaches are used to describe the interaction sign and strength . in case of strong attractive interactions a distinction between the final state scattering and the formation of a baryonic resonance leading to bound or quasi bound states can not be made uniquely . such investigation will be presented.isospin violation or charge symmetry breaking processes are a topical and interesting field of threshold production physics . the essential contribution to isospin breaking is the possible @xmath30 mixing . therefore experiments around the @xmath11 threshold as e.g. @xmath31 or @xmath32 should identify the @xmath30 mixing angle which in turn might give hints to the mass difference of the up and down quarks . threshold production experiments are characterized by excess energies which are small compared to the produced masses discussed in this article , we will principally refer to experimental results and theoretical investigations covering excess energies up to @xmath33 ( the @xmath9mass ) in excess of the respective thresholds . ] . consequently , in fixed target experiments the momenta of the final state particles transverse with respect to the direction of the incoming beam are small compared to the longitudinal components . thus , ejectiles are confined to a narrow forward cone in the laboratory system around the centre of mass movement and close to threshold an experimental acceptance covering the full solid angle is feasible with comparatively small dedicated detector arrangements . small relative momenta in the final state effectively limit the number of partial waves contributing , simplifying the theoretical interpretation of experimental results . it should be noted , that already three body final states require in principle a three body faddeev like approach , which has not been accomplished so far @xcite . however , as first described by watson @xcite , with two strongly interacting particles in the final state , the energy dependence of the total cross section close to threshold is essentially determined by the ( three body ) phase space and the energy dependence of the on shell final state interaction ( fsi ) . due to small relative velocities fsi effects are inherent to the experimental observables . thus , the interpretation of data in terms of reaction dynamics requires a correct treatment of both initial regime , even for proton proton scattering no reliable model exists to allow a consistent evaluation of the initial state interaction @xcite . ] ( isi ) and final state interactions @xcite ( for a review see @xcite ) . on the other hand , fsi effects might provide access to low energy scattering parameters , which are difficult to obtain otherwise in case of unstable particles . meson production at threshold implies high momentum transfers @xmath34 , given by @xmath35 for the reaction type @xmath36 with @xmath37 , @xmath38 and @xmath13 denoting a nucleon , baryon or meson in the initial and final state and @xmath39 , @xmath40 and @xmath41 as the respective masses . with momentum transfers in the range of @xmath42 to @xmath43 for @xmath44 and @xmath45 meson production , threshold production probes with corresponding distances between @xmath46 and @xmath47 the short range part of the nucleon nucleon interaction . consequently , the energy dependence of the primary production amplitude is expected to be weak , motivating watson s approach @xcite . theoretical studies have been carried out mainly within the framework of hadronic meson exchange models , i.e. with baryons and mesons as effective degrees of freedom ( for a recent review see @xcite ) . chiral perturbation theory has been applied for the description of data on @xmath9 production close to threshold @xcite . however , in view of the characteristic distances mentioned above , qcd inspired models @xcite with constituent quarks and gluons or instantons as relevant degrees of freedom , might turn out to be appropriate . high quality exclusive data on close to threshold meson production in the energy range of non perturbative strong interaction physics will be crucial for exploring the boundary between effective meson exchange models and ( so far phenomenological ) approaches based on quark gluon degrees of freedom . it seems that the @xmath6 is of limited value once strangeness is involved in the hadronic systems and non perturbative coupled channel considerations beyond @xmath6 are on the market @xcite . high quality data on the meson ( @xmath48 ) production at threshold have been and are being produced at the hadronic beam cooler rings using the baryon number @xmath49 systems ( essentially @xmath50 and @xmath51 ) . at least equally important are studies with hadronic interactions in the @xmath52 sector . however , most measurements with @xmath9 or @xmath10beams used bubble chamber techniques which naturally suffer from very poor statistics . accurate investigations of differential cross sections for meson production from the initial @xmath9 plus proton or @xmath10 plus proton systems are needed at several momenta to study the s and p wave dynamics and their interplay . there are several indications that as intermediate states both the @xmath53 resonance and the @xmath54 quasi bound state followed by a final state interaction leading to nucleon and meson are essential and could be investigated best via the @xmath52 sector . until now , most of the data for threshold meson production were observed with proton beams . the use of antiprotons has contributed considerably to the knowledge of the meson spectroscopy up to serious candidates for glueballs and exotic hybrids but is still suffering from the lack of unique interpretations . a comparison of the meson production in the fundamental process to productions on heavier nucleons would give constraints to eliminate uncertainties in the basic interactions . as outlined by w. weise and u.g . meiner @xcite , precise measurements of such processes set the necessary constraints for the effective lagrangian of low energy qcd with strange quarks . its detailed knowledge has impacts on several other important issues , such as the strange quark content of the nucleon . though in the present article we will restrict ourselves to hadron induced reactions , in general induced with @xmath55 , @xmath56 , @xmath9 and @xmath57beams , investigations on the structure of the nucleon as seen with leptons should at least be mentioned @xcite . the use of leptonic interactions via photon or electroweak boson exchanges has the attraction of combining a weakly interacting particle of well controlled energy momentum transfer with the complex multi particle system containing charged constituents . photoproduction provides a crucial complementary information compared to hadronic processes , since they are characterized by the electromagnetic probe as a virtual vector meson or an @xmath58 fluctuation and as such are a powerful source of spin1 states . the well proven perturbative methods allow us to define a hierarchy of operators which , at least in principle , can be extracted separately . perturbative methods finally allow to overcome one of the limitations of the electroweak interaction , that is the access to gluon distributions . in fact the strong interaction in the limit of hard scales is becoming a probe for the soft region . all this can be complemented by the addition of the spin and flavour degrees of freedom @xcite . cross sections are very small in the threshold region down to eight orders of magnitude compared to the total yield . due to the rapid growth of the phase space volume the total cross section of the meson production reactions increases by orders of magnitudes in a few mev range of excess energy . such studies have been made possible only due to the low emittance and small momentum spread proton beams available at the storage ring facilities . figure [ dieter_wasserfall ] demonstrates threshold production measurements of light mesons in the proton proton interaction . especially hadron accelerators offer the possibility to excite nuleonic resonances as @xmath59 s and @xmath60 s with the same quantum numbers as those of the proton via the @xmath61 reaction , where the @xmath62channel with its spin and isospin zero serves as a spin isospin filter . this scalar isoscalar tool is complementary to the vector excitation of nuclei with electromagnetic probes . recent results from different electron beam facilities as elsa and mami in germany and nikhef in the netherlands certainly deserve a detailed and circumstantial presentation . due to the present limitation to hadronic interactions we can not give appropriate credit to this field of physics and refer to @xcite . at the continuous electron beam accelerator ( cebaf ) of the thomas jefferson national accelerator facility ( tjnaf ) an electron beam is used for exciting particular nuclear resonances and observing their individual decay modes . the interaction of hadrons being the reflection of the strong force acting between their constituents delivers indirect information about their structure and the strong interaction itself . in the frame of the optical potential model the hadronic interaction can be expressed in terms of phase shifts , which in turn are described by the scattering length and effective range parameters . these are quite well known for the low energy nucleon nucleon interaction @xcite yet they are still poorly established in case of meson nucleon or even meson meson interactions . this is partly due to the absorption of mesons when scattering on a baryon . to account for this effect the scattering length becomes a complex quantity where the imaginary part e.g. in case of the nucleon@xmath11 interaction describes the @xmath63 and @xmath64 processes . moreover , the short life time of all _ neutral _ ground state mesons prohibits their utilization as secondary beams and therefore the study of their interaction with hadrons is accessible only via observations of their influence on the cross section of the reactions in which they were produced ( @xmath65 ) . when created close to the kinematical threshold with the relative kinetic energy being in the order of a few mev , the final state particles remain much longer in the range of the strong interaction than the typical life time of @xmath66 or @xmath20 baryon resonances with @xmath67 and hence they can easily experience a mutual interaction before escaping the area of an influence of the hadronic force . this as introduced by watson @xcite final state interaction ( fsi ) may significantly modify both the original distributions of relative momenta of the outgoing reaction products and the magnitude of the production cross section . though it is easy to understand that the fsi changes the distributions of the differential cross sections it is rather difficult to cope for the influences on the magnitude of the total reaction rate , since one tends to separate the primary production from the final state interaction in space as well as in time @xcite . considering the primary production as a separate process it is well worth trying to understand the phenomenon qualitatively . if there were no final state interactions the total cross section would be fully determined by the kinematically available phase space volume , @xmath68 , where each interval is populated with a probability governed by the primary production amplitude only : @xmath69 the approximation in the equation results from the assumption that @xmath70 in a few mev range above the production threshold @xcite . f denotes the flux factor of the colliding particles . in the classical picture we might imagine that the reaction particles are created together with their appropriate force field and when escaping the interaction region they acquire a potential energy increasing or decreasing their kinetic energy depending whether the interaction is repulsive or attractive . for an attractive interaction they could be created also in those phase space partitions which are not available for non interacting particles and subsequently be `` pulled down '' to the energetically allowed regions by final state interaction . the temporary growth of the primary production phase space would than increase the reaction rate . contrary , in case of a repulsive interaction the particles must be produced in the lower phase space volume since leaving the interaction area they will acquire additional kinetic energy . the reduction of the total cross section , for example in case of the repulsive coulomb force , can easily be understood when considering the production in a coordinate space of point like objects . here in contrast to non interactive particles a strongly repulsive object can not be produced at appropriately small distances since their later acceleration would lead to the breaking of the energy conservation and thus the space available to primary production is reduced . one can also argue relativistically that the primary mechanism creates the particles off the mass shell and subsequently they are lifted onto the mass shell by the final state interaction . the solid line in figure [ dalitz_phasespace ] ( left ) depicts the boundary of the dalitz plot in case of the @xmath71 reaction calculated at the total centre of mass energy @xmath72 exceeding by @xmath73 the threshold energy . the area surrounded by that curve is a direct measure of the kinematically available phase space volume . the dotted line shows the corresponding plot at the moment of the primary creation if the mass of each particle was reduced by @xmath74 , demonstrating that now the available phase space grows significantly . indeed , as shall be inferred from the experimental results presented in subsequent sections , at excess energies of a few mev above threshold , the mutual interaction among the outgoing particles enhances drastically by more than an order of magnitude the total cross section and modifies appreciably the occupation of the phase space . figure [ dalitz_phasespace ] ( right ) indicates the phase space distribution expected for the @xmath75 system at an excess energy of @xmath76 , assuming a homogeneous primary production and taking into account the s wave interaction between the protons . the proton proton fsi modifies the homogeneous dalitz plot distribution of `` non interacting particles '' , enhancing its population at a region where the protons have small relative momenta . the interaction of the proton@xmath11 system would manifest itself at low invariant masses @xmath77 corresponding to small relative momenta between the proton and the @xmath11 meson . such effects observed in the experiments are presented in section [ smpinns ] . at the vicinity of the threshold , relative angular momenta larger than @xmath78 will not play any role due to the short range of the strong interaction and small relative momenta of the produced particles . it can be inferred from parity and angular momentum conservation laws that the production of an @xmath37@xmath37@xmath79 system ( for pseudoscalar or vector mesons ) in relative s waves may only occur if the nucleons collide in p wave . thus at threshold the transition @xmath80 is the only possible one , with capital letters denoting the angular momentum between nucleons and the small letter for the meson angular momentum with respect to the pair of nucleons . an interaction of the nucleons in the entrance channel influences the production process appreciably @xcite , similarly as that described above between the outgoing particles . for example , for @xmath11 production in the @xmath81 reaction the initial state interaction ( isi ) reduces the total cross section by about a factor of 35 @xcite due to the repulsive proton proton @xmath82wave potential . this factor keeps constant in the range of a few tens of mev @xcite and hence , does not influence the energy dependence of the total cross section of the meson production , which remains predominantly determined by the final state interaction . in particular , the dominant s wave nucleon nucleon final state interaction is by far stronger than any of the low energy meson nucleon ones , with the only exception of the proton@xmath83 fsi ( see section [ mwhs ] ) . the effect of the nucleon nucleon fsi diminishes with increasing excess energy since it significantly influences only that partition of the phase space at which the nucleons have small relative momenta . whereas this fraction keeps constant , the full volume of the phase space grows rapidly : an increase of the excess energy from @xmath84 to @xmath85 corresponds to a growth of @xmath68 by more than three orders of magnitude @xcite . as a result the s wave nucleon nucleon fsi is of less importance for higher excess energies where it affects a small fraction of the available phase space volume only . a more quantitative discussion about the influence of the isi and fsi effects at close to threshold production cross sections will be presented in section [ mwhs ] . investigations of the production of mesons and their interactions with nucleons are based on measurements determining the total and differential production cross sections and their dependence on the energy of the interacting nucleons . therefore , to enable a quantitative discussion on the mechanisms leading to the transformation of the energy of motion of nucleons into matter in the form of mesons let us recall the formula of the reaction cross section . in case of the @xmath86 process with the four momenta of the colliding nucleons denoted by @xmath87 and @xmath88 and with @xmath89 particles in the exit channel this reads : @xmath90 where @xmath91 denotes the square of the lorentz invariant spin averaged matrix element describing the probability to create two nucleons and a meson with four momenta of @xmath92 and @xmath93 , respectively . the energy and momentum conservation as well as the on shellness of the created particles is ensured by the dirac@xmath94 and the heaviside@xmath95 functions . the formula holds also for @xmath96 . the total cross section is then defined as an integral of the probabilities to populate a given phase space interval over the whole kinematically available range determined by energy and momentum conservation normalized to the flux factor f of the colliding nucleons . in the case of non interacting final state particles the matrix element close to threshold @xmath91 , is nearly constant @xcite and hence the allowed phase space volume , @xmath68 , is the decisive quantity which governs the growth of the cross section with increasing excess energy q. the latter defined as the total kinetic energy is shared among the outgoing particles in the reaction centre of mass frame : @xmath97 where @xmath98 ) , the total reaction energy @xmath99 amounts to : @xmath100 . before writing explicitly the formula for @xmath68 let us introduce the kinematical triangle function @xmath101 defined as @xcite : @xmath102 which enables the expression of many useful kinematical variables in a very compact and lorentz invariant form . in particular , the momenta of particle @xmath103 and @xmath104 in their rest system equal : @xmath105 where @xmath106 stands for the square of the invariant mass of the @xmath107 system considered as one quasi particle . the above relation gives an expression for the flux factor f from equation in terms of the colliding masses of the nucleons and the total energy s only , namely : @xmath108 where we have chosen the convention introduced by byckling and kajantie @xcite and included the @xmath109 factors for the phase space @xmath110 and for the matrix element @xmath111 into the definition of f. it is important to note that at threshold , for an excess energy range of a few tens of mev , the small fractional changes of total energy @xmath112 causes weak variations of the flux factor and influences only slightly the shape of the energy dependence of the total cross section . for a two particle final state @xmath113 ( for instance for the reaction @xmath114 ) the phase space integral defined in equation reduces to @xmath115 and the total cross section for such reactions ( when neglecting variations due to the dynamical effects @xmath116 ) should increase linearly with the centre of mass momentum of the produced meson in the vicinity of the threshold . the total cross section can be expressed analytically as a function of the masses of the particles participating in the reaction and the square of total reaction energy s : @xmath117 near threshold , at a given excess energy @xmath118 , the emission of the reaction products in the centre of mass frame is isotropic and the whole dynamics of the process manifests itself in the absolute value of the transition matrix element @xmath119 . the underlying production mechanisms can also be extracted from the deviations of the total cross section energy dependence following the prediction of relation . a visualization of possible differences in the dynamics for the production of various mesons can be realized by comparing the total cross section of the studied reactions at the same value of the phase space volume normalized to the flux factor ( @xmath120 ) . it is important to recognize that the comparison at the same centre of mass meson momentum ( @xmath121 ) gives direct information about the amplitude differences only if the production of mesons with the same masses is concerned ( for example @xmath122 and @xmath123 ) , yet the reactions @xmath114 and @xmath124 have by far different @xmath68 at the corresponding @xmath121 . however , if one is interested in the decomposition of the production amplitude according to the angular momenta of the final state particles then indeed the appropriate variable for the comparison even for particles with different masses is the meson momentum in the reaction centre of mass frame . correspondingly for the more than two body final state , the adequate variable is the maximum meson momentum , since it is directly connected to the maximum angular momentum by the interaction range . considering for a three body exit channel ( @xmath125 ) that the meson ( @xmath126 ) possesses maximum momentum when the remaining two particles are at rest relative to each other and employing definition one obtains : @xmath127 contrary to the two body final state , at a fixed excess energy the dynamics of the meson production associated with two or more particles reflects itself not only in the absolute value of the square of the matrix element but also in distributions of variables determining the final state kinematics . usually , in non relativistic calculations of the total cross section , one takes the jacobi momenta , choosing as independent variables the qmeson momentum in the reaction centre of mass frame and kmomentum of either nucleon in the rest frame of the nucleon nucleon subsystem . by means of the @xmath101 function they can be expressed as : @xmath128 with @xmath129 denoting the square of the invariant mass of the nucleon nucleon subsystem . in a non relativistic approximation the expression of the total cross section defined by formula for a meson production reaction in nucleon nucleon interactions of the type @xmath130 simplifies to : @xmath131 denoting by @xmath132 and @xmath133 the relative angular momentum of the nucleon nucleon pair and of the meson relative to the @xmath12 system , respectively and approximating the final state particles by non distorted plane waves ( case of non interacting objects ) , whose radial parts @xmath134 are given by the spherical bessel functions : @xmath135 an expansion of the transition amplitude @xmath136 for the @xmath137 partial wave around @xmath138 leads to @xmath139 applying the above proportionality in equation and solving the integral , yields the partial cross section @xmath140 where @xmath141 with @xmath142 denoting the mass of the created meson , here in order to avoid ambiguities with the abbreviation for the eta - meson , we introduce an additional suffix @xmath5 . ] . thus at threshold for the ss partial wave the cross section for the @xmath86 reaction should increase with energy of the fourth power of @xmath143 . the dimensionless parameter @xmath143 was introduced by rosenfeld @xcite as a variable allowing for the qualitative estimations of the final state partial waves involved in pion production . he argued that if the phenomenon of pion production takes place at a characteristic distance r from the centre of the collision , with r in the order of @xmath144 , then the angular momentum of the produced meson is equal to @xmath145 . hence , @xmath146 denotes the classically calculated angular maximum momentum of the meson relative to the centre of reaction . the same arguments one finds in the work of gell mann and watson @xcite , where the authors do not expect the range of interaction to be greater than the compton wave - length of the produced meson @xmath147 . however , it is rather a momentum transfer @xmath34 between the colliding nucleons which determines the distance to the centre of the collision @xmath148 at which the production occurs . based on the data indications , it is emphasized in the original article of rosenfeld @xcite that r is slightly less than @xmath149 , which is numerically close to the value of @xmath150 . directly at threshold , where all ejectiles are at rest in the centre of mass frame , @xmath34 is equal to the centre of mass momentum of the interacting nucleons and hence , exploring equation it can be expressed as : @xmath151 though the present considerations are constrained to the spin averaged production only , it is worth noting that very close to threshold due to the conservation laws and the pauli excluding principle , where l , s and t denote angular momentum , spin and isospin of the nucleon pair , respectively . for example if the nucleon nucleon wave function is symmetric in space ( @xmath152 ) and spin ( @xmath153 ) then it must be antisymmetric in isospin space ( @xmath154 ) to be totally antisymmetric . ] for many reactions there is only one possible angular momentum and spin orientation for the incoming and outgoing particles . -0.25 cm = 1em plus2em minus.5em to for example the production of neutral mesons with negative parity as pseudoscalar or vector mesons may proceed in the proton proton collision near threshold only via the transition between @xmath155 and @xmath156 partial waves ( see table [ partialtransitions ] ) . therefore only the collision of protons with relative angular momentum equal to @xmath157 and a parallel spin orientation which flips during the reaction may lead to the production of negative parity neutral mesons . in the case of the production of neutral scalar mesons , the protons or neutrons must collide with anti parallel spin orientations which remains unchanged after the reaction . these simple considerations imply that in the close to threshold measurements with the polarized beam and target one should see a drastic effect in the reaction yield depending whether the polarization of reacting protons is parallel or anti - parallel . in the ideal case , having the product of beam and target polarization equal to unity , the close to threshold @xmath44 , @xmath11 or @xmath158 production can be realized only in case of the parallel spin orientation of the reacting protons and for the anti - parallel polarizations zero yield should be observed . indeed , in reality , strong differences in the yield for various combinations of beam and target polarization have been determined in the pioneering measurements of the reaction @xmath159 at the iucf facility @xcite . due to both the strong nucleon nucleon low energy interaction in the @xmath16 state and the meson nucleon forces in the exit channel , the assumption of the non interactive plane waves leading to equation failed when confronted with experimental data . however , these deviations give the possibility for determining the still poorly known nucleon meson interactions . the interaction between particles depends on their relative momenta . consequently , for investigations of final state interactions more instructive coordinates , than the q and k momenta , are the squared invariant masses of the two body subsystems @xcite . these are the coordinates of the dalitz plot . in the original paper @xcite dalitz has proposed a representation for the energy partitions of three bodies in an equilateral triangle whose sides are the axes of the centre of mass energies . he took advantage of the fact that the sum of distances from a point within the triangle to its sides is a constant equal to the height . therefore , the height of the triangle measures the total energy @xmath160 and interior points fulfilling four momentum conservation represent energy partitions . for a constant @xmath99 due to the energy conservation , without loosing any information , it is enough to consider the projection on any of the @xmath161 planes . the linear relation between @xmath162 and @xmath163 ( @xmath164 ) allows to use @xmath165 or @xmath161 coordinates equivalently , with the following relation between the phase space intervals : @xmath166 . according to the kilian s geometrical representation @xcite a dalitz lies on a plane in the three dimensional space ( @xmath167 ) orthogonal to the space diagonal . the plane including a dalitz plot corresponding to a given total energy s is then given by the following scalar product @xcite : @xmath168 the second equality of equation means that there are only two independent invariant masses of the three subsystems and therefore a projection onto any of the @xmath169 planes still comprises the whole principally accessible information about the final state interaction of the three particle system . in the case of no dynamics whatsoever and the absence of any final state interaction the occupation of the dalitz plot would be fully homogeneous since each phase space interval would be equally probable . the final state interaction would then appear as a structure on that area . figure [ dalitz_examples]a shows for the example of the @xmath170 reaction how the uniformly populated phase space density is modified by the s wave ( @xmath16 ) interaction between outgoing protons . an enhancement in the range corresponding to the low relative momentum between protons is clearly visible . a steep decrease of the occupation density with increasing invariant mass of the proton proton subsystem is even better seen in figure [ dalitz_examples]b presenting the projection of the phase space distribution onto the @xmath171 axis indicated by an arrow in figure [ dalitz_examples]a . this is a direct reflection of the shape of the proton proton ( @xmath16 ) partial wave amplitude shown in figure [ dalitz_examples]c . figure [ dalitz_examples]d shows the dalitz plot distribution simulated when switching off the proton proton interaction but accounting for the interaction between the @xmath11meson and the proton . due to the lower strength of this interaction the expected deviations from the uniform distributions are by about two orders of magnitude smaller , but one recognizes a slight enhancement of the density in the range of low invariant masses of proton@xmath11 subsystems . however , due to weak variations of the proton@xmath11 scattering amplitude the enhancement originating from the @xmath11meson interaction with one proton is not separated from the @xmath11meson interaction with the second proton . therefore an overlapping of broad structures occurs . it is observed that the occupation density grows slowly with increasing @xmath171 opposite to the effects caused by the s wave proton proton interaction , yet similar to the modifications expected for the p wave one @xcite . from the above example it is obvious that only in high statistics experiments signals from the meson nucleon interaction can appear over the overwhelming nucleon nucleon final state interaction . it is worth noting , however , that the dalitz plot does not reflect any possible correlations between the entrance and exit channel @xcite . the dalitz plot representation allows also for a simple interpretation of the kinematically available phase space volume as an area of that plot . namely , equation becomes : @xmath172 where the limits of integrations defining the boundaries of the dalitz plot can be expressed as @xcite : @xmath173 thus , the phase space volume kinematically available for the three body final state can be written by means of only one integral : @xmath174 whose solution leads , in general , to elliptic functions @xcite . however , in the nonrelativistic approximation it has the following closed form : @xmath175 where the substitution of the non relativistic relation between @xmath176 and q @xmath177 gives the @xmath178 partial cross section of equation . on the basis of the formula the kinematically available phase space volume ( @xmath68 ) can be as easily calculated as the excess energy q. close to threshold in the range of a few tens of mev the non relativistic approximation differs only by a few per cent from the full solution given in equation , which in fact with an up to date computer can be solved numerically with little effort . therefore , in the following chapters we will describe the data as a function of @xmath68 as well as of q or @xmath143 , if it is found to be appropriate . the choice of the proper observables for picking up the dynamical effects in the case of a four particle exit channel is by far more complicated . it was for example discussed by chodrow @xcite who has given the extension of the dalitz representation to the four body final states . he has proven that the two body correlation in the case where there are two identical particles in the four body final state say @xmath107 can be directly visible in the density distribution of events on the @xmath179plane , where @xmath180 denotes the energy of particle @xmath103 and @xmath181 is defined as : @xmath182 extending meson production experiments from the free nucleon nucleon interaction to interactions of light nuclei is interesting for many reasons . first of all it offers the possibility to study more complex production processes with more than only one target nucleon involved . furthermore , in the absence of free neutron targets , it gives the possibility to study the meson creation in the proton neutron interaction at proton beam facilities , complementary to a neutron induced @xmath9 meson production with use of high resolution ( energy spread less than @xmath183 ) secondary neutron beams @xcite investigated at triumf @xcite . the technique based on quasi free proton nuclear reactions , with a beam of protons reacting with a nucleus in the target is utilized at the celsius and cosy accelerators @xcite . first experiments of @xmath184 meson production in the proton neutron reaction with simultaneous tagging of the spectator proton have recently been carried out at celsius and resulted in a resolution of the excess energy equal to @xmath185 @xcite . in the quasi - free approximation ( fig . [ qfree ] ) it is assumed that the bombarding proton interacts exclusively with one nucleon in the target nucleus and that the other nucleons affect the reaction by providing a momentum distribution to the struck constituent only . this assumption is justified if the kinetic energy of a projectile is large compared to the binding energy of the hit nucleus . in fact , as noticed by slobodrian @xcite , also the scattering of protons on a hydrogen target , where the protons are bound by molecular forces , may serve as an extreme example of the quasi free reaction . in that case , although the hydrogen atoms rotate or vibrate in the molecule , their velocities and binding forces are totally negligible with respect to the velocity and nuclear forces operating the scattering of the relativistic proton @xcite . the deuteron is also relatively weakly bound with a binding energy of @xmath186 , which is by far smaller more than two orders of magnitude in the case of pion and already more than three orders of magnitude in the case of @xmath45 meson production than the kinetic energy of the bombarding protons needed for the creation of mesons in the proton neutron interaction . however , it has to be considered that even the low binding energy of @xmath186 results in large fermi momenta of the nucleons which can not be neglected . the momentum and kinetic energy spectra of the nucleons in the deuteron are shown in figure [ fermi_mom_and_kin ] . in the considered approximation the fermi motion of the nucleons is the only influence of the internucleon forces on the proton neutron reaction and hence the struck neutron is treated as a free particle @xmath187 in the sense that the matrix element for quasi free meson production off a bound neutron is identical to that for the free @xmath188 reaction . the proton from the deuteron is considered as a spectator which does not interact with the bombarding particle , but rather escapes untouched and hits the detectors carrying the fermi momentum possessed at the moment of the collision . from the measurement of the momentum vector of the spectator proton @xmath189 one can infer the momentum vector of the neutron @xmath190 at the time of the reaction and hence calculate the excess energy q for each event , provided that the beam momentum is known . as an example , a distribution of the excess energy in a quasi free @xmath191 reaction is presented in figure [ q_and_mass_off]a @xcite . due to the large centre of mass velocity ( @xmath192 ) with respect to the colliding nucleons , a few mev wide spectrum of the neutron kinetic energy inside a deuteron ( fig . [ fermi_mom_and_kin]b ) is spread by more than a factor of 30 . + therefore , especially in the case of near threshold measurements , where the cross section grows rapidly with increasing excess energy ( see e.g. fig . [ dieter_wasserfall ] ) the total centre of mass energy @xmath99 has to be determined on an event by event level . for this purpose , in experiments , the spectator protons are usually registered by silicon pad or @xmath193strip detectors @xcite which determine their kinetic energy ( @xmath194 ) and polar emission angle ( @xmath195 ) . thus , it is useful to express s as function of these variables : @xmath196 with @xmath197 denoting the centre of mass energy , assuming a vanishing fermi motion . measuring both the energy and the emission angle of the spectator protons it is possible to study the energy dependence of a meson production cross section from data taken at only one fixed beam momentum . it must be noted , however , that in the framework of the impulse approximation , illustrated in figure [ qfree ] , the measured spectator proton is a physical particle , yet the reacting neutron is off its mass shell , where the explicit expression for its four momentum vector @xmath198 , in the rest frame of the deuteron , reads : @xmath199 with @xmath194 and @xmath189 denoting the kinetic energy and the momentum vector of a spectator proton , respectively . the mass spectrum of the interacting neutron ( @xmath200b . it can be seen that the maximum of this spectrum differs only by about @xmath201 from the free neutron mass ( @xmath202 ) , however on the average it is off by about @xmath203 . in the frame of the discussed approximation , the struck neutron is never on its mass shell and the minimum deviation from the real mass occurs for vanishing fermi momentum and as can be inferred from equation is equal to the binding energy @xmath204 . measurements performed at the celsius and triumf accelerators for the @xmath71 @xcite and @xmath205 @xcite reactions , respectively , have shown that within the statistical errors there is no difference between the total cross section of the free and quasi free processes . in figure [ free_vs_quasifree]a the production of the @xmath11 meson in free proton proton collisions is compared to the production inside a deuteron and in the overlapping regions the data agree within the statistical errors . these observations allow to anticipate that indeed the assumption of the identity for the transition matrix element for the meson production off free and quasi free nucleons bound in the deuteron is correct . in case of the meson production off the deuteron , one can also geometrically justify the assumption of the quasi free scattering since the average distance between the proton and the neutron is in the order of @xcite . ] @xmath206 . of course , the other nucleon may scatter the incoming proton and the outgoing meson . however , these nuclear phenomena are rather of minor importance in case of the production on the neutron bound in the deuteron , but should be taken into account for derivations of total cross sections from experimental data . the reduction of the beam flux on a neutron , due to the presence of the proton , referred to as a shadow effect , decreases for example the total cross section by about @xmath207 @xcite for the @xmath11meson production . similarly , the reduction of the total cross section due to the reabsorption of the outgoing @xmath11 meson on the spectator proton was found to be only about @xmath208 @xcite . the appraisals were performed according to a formula derived in reference @xcite which shows that the cross section for the deuteron reduces by a factor of : @xmath209 compared to the free nucleon cross sections . here @xmath210 denotes the @xmath211 inelastic cross section and @xmath212 stands for the average of the inverse square nucleon separation in the deuteron taking the nucleon size into account @xcite . the latter effect for the production of mesons like @xmath213 is expected to be much smaller , since ( as shall be discussed in section [ mwhs ] ) the s wave interaction of the @xmath11meson with nucleons is by far stronger than for any of the mentioned ones . due to the relatively large distance between the nucleons in the deuteron one could argue that the quasi free reactions will dominate the production of mesons even with a deuteron in the final state . the conviction that the production of mesons in nucleon deuteron collisions proceeds predominantly via an elementary scattering of nucleons is also based on the comparative investigations of the @xmath44meson production in proton neutron and proton deuteron collisions @xcite . the total cross section for the @xmath214 reaction close to the kinematical threshold is by more than two orders of magnitude lower compared to the elementary @xmath124 process ( see figure [ free_vs_quasifree]b ) . this observation has been successfully explained assuming that the production of the @xmath44 meson in the @xmath214 reaction is proceeding entirely due to the quasi free @xmath124 reaction @xcite for @xmath143 larger than 0.7 ( dashed line ) . in the threshold region the result of the calculation ( solid line ) describes exactly the experimental data when the proton deuteron fsi is additionally taken into account . the large difference in magnitude between the threshold cross section for the @xmath124 and the @xmath214 reaction can be explained since the incident proton momentum needed to create a @xmath44 meson via the @xmath214 process is much below the kinematical threshold for the production in elementary proton neutron collisions . therefore only nucleons possessing large fermi momenta can contribute to the @xmath44 production and this appears with a small probability ( compare fig . [ fermi_mom_and_kin]a ) . the other possible elementary processes @xmath215 and @xmath216 can be neglected since they are expected to lead predominantly to a four body exit channel . this explanation could serve as a heuristic example , however recent data @xcite on the cross sections of @xmath217 show that the supposed spectator protons to the elementary @xmath218 reaction are observed in phase space regions corresponding to zero fermi momenta and hence the quasi free production could be significant only at energies close to the @xmath17 threshold ( @xmath219 ) . in this experiment the deuteron beam was impinging onto the hydrogen target , and hence the spectator protons were boosted in the laboratory permitting registration also of the low energy part of the spectators spectral distribution . a natural extension of the meson production in quasi free proton neutron collisions with either a deuteron used as the target or as the beam is the combination of both methods into a double quasi free production occurring during the collisions of deuterons as depicted in figure [ dd_ppnneta ] . utilizing this technique @xcite and registering both the slow and fast spectator protons could allow for the study of meson production in neutron neutron collisions with a very precise determination of the excess energy which depends on the accuracy for the registration of the momentum vectors of the spectators . the simplest method to study the interactions between meson and nucleon is to scatter a beam of mesons off hydrogen or deuteron targets and to measure the behaviour of the total and differential cross section in the elastic scattering of the considered particles . figure [ kpkp_kk_sspspp]a presents the total cross section for @xmath220 and @xmath221 elastic scattering as a function of the centre of mass excess energy . when comparing the data to calculations @xcite including the changes of phase space integral ( eq . ) and coulomb interaction in the initial and final states one observes a huge enhancement for the @xmath221 cross section with decreasing excess energy and a slight suppression in the case of @xmath220 scattering . this observation may be attributed to the slight repulsion due to the kaon proton hadronic interaction and the significantly larger scattering caused by the strong interaction between the @xmath83 and proton due to the vicinity of the @xmath222 hyperon resonance . the effect has a direct influence on the effective mass splitting of kaons immersed in dense nuclear matter @xcite . a coulomb repulsion was taken into account by multiplying the expression by the coulomb penetration factor @xmath223 for both exit and entrance channels . @xmath223 determines the ratio of the probability of finding two particles close together to the probability of finding two uncharged particles together , all other things being equal @xcite and can be expressed as @xcite : @xmath224 where @xmath225 is the relativistic coulomb parameter , which for the collision of particles @xmath226 reads : @xmath227 with the fine structure constant @xmath62 , the relative velocity v of the colliding particles and with @xmath228 , @xmath229 denoting their charges larger than @xmath230 the @xmath223 of equation needs to be multiplied by a factor of @xmath231 @xcite . ] . the comparison of the @xmath232p scattering data with the phase space integral presented in figure [ kpkp_kk_sspspp]a aims only to illustrate the qualitative influence of the meson - nucleon dynamics on the total cross section energy dependence . it is worth noting , however , that at present the data can be qualitatively well described in the frame of the effective chiral lagrangian with a coupled channel potential @xcite . another method permitting such investigations is the production of a meson in the nucleon nucleon interaction close to the kinematical threshold where the outgoing particles possess small relative velocities . the experimental observables are the excitation functions and the double differential cross section . a possible deviation from the expectation based on the assumption of homogeneous phase space abundance corrected for the known nucleon nucleon interaction delivers information about the meson nucleon forces . for example , figure [ kpkp_kk_sspspp]b presents the first data points on the close to threshold production of the @xmath83 meson via the @xmath233 reaction . the superimposed solid line depicts the results of calculations @xcite taking into account the changes of the production amplitude as deduced from the @xmath220 and @xmath221 elastic scattering shown in figure [ kpkp_kk_sspspp]a , but neglecting the influence of the dominant proton proton interaction ! on the other hand the proton proton fsi can not be neglected in case of three body final states , e.g. @xmath234 @xcite , @xmath170 @xcite , or @xmath235 @xcite , since it has strong influence on the total cross section energy dependence enhancing it by more than an order of magnitude for excess energies below @xmath236 as shall be shown below . thus it is surprising that in spite of its neglection one can describe the data of the @xmath237 reaction . the origin of that effect will be investigated experimentally in the near future @xcite . at present one can only speculate whether it is due to the partial compensation of the @xmath50 and @xmath221 hadronic interaction or maybe due to the additional degree of freedom given by the four body final state @xcite . in order to demonstrate the influence of the nucleon nucleon interaction on the cross section dependence on the excess energy let us consider the production of neutral pseudoscalar mesons ( @xmath44 , @xmath11 , @xmath158 ) in the collision of protons . due to the short life time of these mesons the study of their low energy interaction with nucleons in direct scattering experiments is not feasible . in general for a three body exit channel one expects a dependence of the total cross section on the energy which can be described by the linear combination of partial cross sections of equation . therefore , to be able to extract the information about the final state interaction of the outgoing particles from the energy dependence it is necessary to know precisely the contribution originating from different partial waves . yet appropriately close to threshold there is only one important combination of the angular momenta of emitted particles ( ss ) and in this region the energy dependence of the total cross section is uniquely determined . however , the range of excess energies for the s wave dominance changes strongly with the mass of the produced meson . for the production of the @xmath44 meson in proton proton collisions the investigations with polarized beam and targets @xcite allowed to deduce that the ss partial wave accounts for more than @xmath238 of the total cross section up to @xmath239 , as can be seen in figure [ kpkp_kk_sspspp]c , where the ss contribution is indicated by the dotted line . the ss contribution was inferred assuming the @xmath240 and @xmath241 dependence for ps and pp partial waves , respectively . these are power laws taken from proportionality , which was derived under the assumption of non interacting particles . relatively small values of @xmath242wave nucleon nucleon phase shifts at low energies ( compared to @xmath16 phase shifts in figure [ 1s03p0_eta_q_vps]a ) and similarly weak low energy interactions of p wave protons in other spin combinations @xcite justify this assumption . the measurements with the polarized beam and target allow a model independent determination of the contribution from individual partial waves . in case of the @xmath159 reaction , due to the identical particles in the initial state and the rotational equivalences @xcite there are only seven independent polarization observables , e.g. two beam analyzing powers and five linear combination of spin correlation coefficients . in particular , as shown in references @xcite the close to threshold contributions of the @xmath243 and @xmath244 partial waves can be determined in the model - free way from the measurements of the spin dependent total cross sections only . for example the strength of the @xmath243 final state can be expressed as @xcite : @xmath245 where @xmath246 denotes the total unpolarized cross section and @xmath247 and @xmath248 stand for differences between the total cross sections measured with anti - parallel and parallel beam and target polarizations . subscripts @xmath249 and @xmath132 associate the measurements with the transverse and longitudinal polarizations , respectively . in accordance with the phenomenology of gell mann and watson @xcite described in the previous section one expects that also in the case of heavier mesons the ss partial wave combination will constitute the overwhelming fraction of the total production cross section for @xmath146 smaller than 0.4 . this implies as can be deduced from the relation between @xmath146 and q illustrated in figure [ 1s03p0_eta_q_vps]b that mesons heavier than the pion are produced exclusively via the ss state in a much larger excess energy range and hence larger phase space volume ( see figure [ 1s03p0_eta_q_vps]c ) . in particular in case of @xmath158 this is larger by more than one order of magnitude . thus , whereas in case of @xmath44 the onset of higher partial waves is observed at q around @xmath73 it is expected only above @xmath250 and above @xmath251 for @xmath158 and @xmath11 mesons , respectively . figures [ differential]b and [ differential]c present the angular distributions of the created @xmath11 meson in proton collisions . it is evident that at @xmath252 and still at @xmath253 the production of the @xmath11 meson is completely isotropic within the shown statistical errors . although at @xmath254 the accuracy of the data does not exclude a few per cent of the contribution originating from higher partial waves , the dominance of the s wave creation is evident . similarly , the measurements of the differential cross section ( fig . [ differential]a ) for the @xmath255 reaction performed at saturne @xcite at @xmath256 are still consistent with pure ss wave production , though the relatively large error bars would allow for other contributions on a few per cent level ( @xmath257 ) . -0.25 cm = 1em plus2em minus.5em to let us now consider to what extent the energy dependence of the total cross section in the estimated range of the dominance of the ss partial waves can be understood in terms of the phase space variation and the interaction between the particles participating in the reaction . watson @xcite and migdal @xcite have proposed the factorization of the amplitude in the case where the production is of short range and the interaction among the outgoing particles is of long range . this requirement is well fulfilled in the case of the close to threshold meson production due to the large momentum transfer ( @xmath34 ) between the interacting nucleons needed to create the considered mesons ( @xmath258 ) . according to the heisenberg uncertainty relation the large momentum transfer brings about a small space in which the primary creation of the meson takes place . in table [ momtranstable ] the distance probed by the @xmath259 reaction at threshold is listed for particular mesons . it ranges from @xmath46 for pion production to @xmath47 for the @xmath45 meson , whereas the typical range of the strong nucleon nucleon interaction at low energies determined by the pion exchange may exceed a distance of few fermi and hence is by one order of magnitude larger than the values listed in table [ momtranstable ] . thus in analogy to the watson migdal approximation for two body processes @xcite the complete transition matrix element of equation may be factorized approximately as @xmath260 represents the total short range production amplitude , @xmath261 describes the elastic interaction among particles in the exit channel and @xmath262 denotes the reduction factor accounting for the interaction of the colliding protons . further , in the first order approximation one assumes that the particles are produced on their mass shell and that the created meson does not interact with nucleons . this assumption implies that the @xmath263 term can be substituted by the square of the on shell amplitude of the nucleon nucleon elastic scattering : @xmath264 proton proton partial wave relevant for further consideration can be expressed explicitly as follows @xcite : @xmath265 where @xmath266 denotes the square root of the coulomb penetration factor defined by equation , k stands for either proton momentum in the proton proton rest frame and the phase shift is indicated by @xmath267 . the phase shifts @xmath268 can be extracted from the said data base ( see fig . [ 1s03p0_eta_q_vps]a ) or , alternatively , can be calculated according to the modified cini stanghellini formula including the wong noyes coulomb correction @xcite + @xmath269 where @xmath270 @xcite , with @xmath225 the same as in eq . . + the phenomenological quantities @xmath271 and @xmath272 denote the scattering length and effective range @xcite , respectively . the parameters @xmath273 and @xmath274 are related to the detailed shape of the nuclear potential and derived from a one pion exchange model @xcite . substituting equation into equation allows to calculate the low energy amplitude for the proton proton elastic scattering @xcite . @xmath275 is taken to be constant for larger values of @xmath276 . ] : + @xmath277 the result is presented as a solid line in figure [ mpppp_cross_pi]a and is in good agreement with the values obtained from the phase shifts of the vpi partial wave analysis @xcite , shown as solid circles and with the phase shifts of the nijmegen analysis @xcite , shown as open squares . the factor @xmath223 is always less than unity due to the coulomb repulsion between protons . at higher energies , where @xmath223 is close to unity , the nuclear scattering will be predominant and for the very low energies the coulomb and nuclear interactions are competing . the coulomb scattering dominates approximately up to about @xmath278 of the proton energy in the rest frame of the other proton , where @xmath223 equals to one half @xcite . in the case of the @xmath279 reaction the maximum possible energy of a proton seen from another proton is equal to @xmath278 already at an excess energy of about @xmath280 . therefore , a significant influence of the coulomb repulsion on the energy dependence of the total production cross section is expected only at very low excess energies , i.e. conservatively for @xmath281 . assuming that the on shell proton proton amplitude exclusively determines the phase space population one can obtain the total cross section energy dependence substituting equation into the formula ( [ vpsdalitz ] ) . solid lines in figures [ mpppp_cross_pi]b and [ cross_etap_pkl]a represent the determined dependence for the @xmath215 and @xmath255 reactions , respectively . the absolute scale of the calculations has been fixed by the normalization to the data . one recognizes the good agreement with the experimental points in the excess energy range up to @xmath282 for @xmath44 meson production and up to @xmath283 for @xmath158 . this confirms the earlier considerations that the domain of ss partial wave dominance should widen significantly with the mass of the created meson and as expected it extends above @xmath250 in case of the @xmath158 meson . comparing the data to the arbitrarily normalized phase space integral of equation reveals that the proton proton fsi enhanced the total cross section by more than an order of magnitude for low excess energies . this is particularly clear in figure [ cross_etap_pkl]a where the logarithmic scale is utilized . however , the much weaker nucleon hyperon interaction increases the total cross section for the close to threshold @xmath284 meson production only by about a factor of two . again in that case as shown in figure [ cross_etap_pkl]b the factorization of the matrix element into constant on shell primary production amplitude and the elastic scattering between outgoing baryons leads to the good description of the close to threshold data . in this case the coulomb interaction between @xmath284 meson and proton was taken into account multiplying the scattering amplitude by the coulomb penetration factor . the dotted line in figures [ mpppp_cross_pi]b and [ cross_etap_pkl]a , which for @xmath285 is practically indistinguishable from the solid line , presents the excess energy dependence of the total cross section taking into account the proton proton fsi effects according to the model developed by fldt and wilkin @xcite . representing the scattering wave function in terms of bound state wave function the authors derived a closed formula which describes the effects of the nucleon nucleon fsi as a function of the excess energy q only . this approach is specifically useful for the description of the spin triplet proton neutron fsi , due to the existence of a bound state ( deuteron ) with the same quantum numbers . though the bound state of the proton proton system does not exist , the model allows also to express the total cross section energy dependence for a @xmath286 reaction by a simple and easily utilizable formula : @xmath287 where the parameter @xmath288 has to be settled from the data . the flux factor f and the phase space volume @xmath68 are given by equations and , respectively . the normalization can be determined from the fit of the data which must be performed for each reaction separately . in the case of @xmath11 meson production the interaction between nucleons is evidently not sufficient to describe the increase of the total cross section for very low and very high excess energies , as can be seen in figure [ cross_eta_fig]a . the solid line in the figure was normalized to the data at an excess energy range between @xmath289 and @xmath290 . the increase of the total cross section for higher energies can be assigned to the outset of higher partial waves . as expected from the previous considerations , this is indeed seen at @xmath291 where the energy dependence of the total cross section starts to change its shape . on the contrary , the close to threshold enhancement being by about factor of two larger than in the case of the @xmath44 and @xmath158 mesons can be assigned neither to the contribution from other than ss partial waves nor to the variation of the primary production amplitude @xmath292 . the latter is expected to change at most by a few per cent for excess energies below @xmath293 @xcite . instead , this discrepancy can be plausibly explained by the influence of the attractive interaction between the @xmath11 meson and the proton . note , that the real part of the scattering length of the @xmath11proton potential depending on the analysis method and the studied reaction is 3 to 10 times @xcite larger than the scattering length for @xmath44proton scattering ( @xmath294 ) @xcite . hence , the modifications of the total cross section energy dependence due to @xmath44proton interaction are too weak to be observed within the up to date accuracy of measurements and calculations . however , the influence of the @xmath11proton interaction presented in figure [ cross_eta_fig ] is evident and hereafter it will be considered whether it may serve for the estimation of the @xmath11proton scattering parameters . the strict quantitative calculation requires , however , the evaluation of the three body faddeev equation . here , we will rather present a simple phenomenological treatment which shall lead to the qualitative understanding how the mutual interaction among three outgoing particles affects the total cross section dependence on excess energy . one of the simplest possibilities based on the naive probabilistic interpretation of the incoherent pairwise interaction would be to factorize the overall enhancement factor into corresponding pair interactions @xcite : @xmath295 where @xmath296 denotes the square of the elastic scattering amplitude of particles @xmath103 and @xmath104 . the @xmath297 can be evaluated according to the formula , which for the s wave @xmath11proton scattering , after substitution of @xmath298 and @xmath299 , reduces to from equation was changed because the imaginary part of the proton@xmath11 scattering length is positive @xcite , as we also adopted here , whereas in the majority of works concerning nucleon nucleon interaction , the scattering length is negative @xcite . we are grateful for this remark to a. gasparyan . ] : @xmath300 where @xmath301 denotes the @xmath11 momentum in the proton@xmath11 rest frame . the scattering length @xmath302 and effective range @xmath303 are complex variables with the imaginary part responsible e.g. for the @xmath304 conversion . the factorization of both i ) the overall production matrix element ( eq . ) and ii ) the three particle final state interactions ( eq . ) applied to the formula gives the following expression for the total cross section of the @xmath71 reaction : @xmath305 where the protons are distinguished by subscripts . exploring formulas and gives the results shown as the dotted line in figure [ cross_eta_fig]a . evidently , the inclusion of the proton@xmath11 interaction enhances the total cross section close to threshold by about a factor of 1.5 and leads to a better description of the data . the effect of the proton@xmath11 interaction is also seen in the experimental distribution of the differential cross sections @xmath306 shown in figures [ cross_eta_fig]b and [ cross_eta_fig]c . these distributions originating from kinematically complete measurements comprise the whole experimentally available information about the interactions of the @xmath307system . in figure [ cross_eta_fig]b one recognizes the increase of the distribution density at regions of small invariant masses of the proton proton and proton@xmath11 subsystems . at this excess energy ( @xmath308 ) these regions are quite well separated . however , since this is close to the energy where the advent of higher partial waves is awaited , the possible contribution from the p wave proton proton interaction can not be a priori excluded . specifically , the latter leads to the enhancement at large invariant masses of the proton proton pair @xcite and hence affects the phase space region where the modification from the proton@xmath11 interaction is expected . figure [ cross_eta_fig]c presents the high statistics data taken at an excess energy of @xmath252 , where the assumption of the total dominance of the ss partial wave is rather save . at this excess energy , corresponding to the small relative momentum range @xmath309 , the variations of the proton@xmath11 scattering amplitude are quite moderate ( fig . [ ratio_and_factors]c ) and as expected from simulations presented in the previous section , the increase of the phase space population due to the @xmath11meson interaction with one proton is not separated from the other one . interestingly , the density growth with increasing invariant mass of the proton proton system is much faster than expected from the simulations presented in figure [ dalitz_examples]d , which have been performed with a scattering length equal to @xmath310 . however , when reducing the proton proton fsi effect to a multiplicative factor , one finds that it depends on the assumed nucleon nucleon potential and on the produced meson mass @xcite . this issue was recently vigorously investigated e.g. by authors of references @xcite and we shall briefly report this here as well . up to now we factorized the transition matrix element into a primary production of particles and its on shell rescattering in the exit channel ( eqs . ) . though it is a crude approximation , neglecting the off shell effects of the production process completely , it astoundingly leads to a good description of the energy dependence of the total cross section , as already demonstrated in figures [ mpppp_cross_pi]b , [ cross_etap_pkl]a , b and [ cross_eta_fig]a . the off shell effects , as pointed out by kleefeld @xcite , could have been safely neglected in case of the electromagnetic transitions in atoms or @xmath311 decays , where the excitation energy of the involved nucleons is by many orders of magnitude smaller than their masses and the initial and final states go hardly off shell @xcite . however , in case of the @xmath130 process the large excitation energy of the colliding nucleons is comparable with the nucleon masses and the primary interaction may create the particles significantly far from their physical masses , so that the off shell effects can not be a priori disregarded . generally , the decomposition of the total production amplitude into the primary production and the subsequent nucleon nucleon interaction visualized in figure [ loop ] is expressed by the formula : @xmath312 where the second term of the equation represents the integration over the intermediate ( @xmath313 ) momenta of the off shell production amplitude and the half off shell nucleon nucleon @xmath249 matrix @xcite . assuming that the primary production occurs in such a way that one nucleon emits the meson which then re scatters on the other nucleon and appears as a real particle the authors of reference @xcite found that the enhancement of the cross section due to the nucleon nucleon interaction depends strongly on the mass of the created meson . this is because with the increasing mass of the produced meson the distance probed by the nucleon nucleon interaction decreases ( see table [ momtranstable ] ) and hence the relevant range of the off shell momenta becomes larger . the effect for the @xmath279 reactions is presented in figure [ barufsi]a , where one can see that , when utilizing the bonn potential model for the nucleon nucleon @xmath249 matrix , the enhancement in case of the @xmath44 production is by about a factor of four larger than for the @xmath11 or @xmath158 mesons . a similar conclusion , but with the absolute values larger by about @xmath314 , was drawn for the paris @xmath12 potential @xcite . on the contrary , when applying the yamaguchi potential into calculations the enhancement grows with the increasing mass of the meson , as shown in figure [ barufsi]b . thick solid curves in figures [ barufsi]a and [ barufsi]b show the results of the frequently applied approximation of the nucleon nucleon fsi effects : @xmath315 where the overall transition matrix element @xmath5 is factorized to the primary on shell production and the @xmath12 fsi expressed as the inverse of the jost function @xmath316 @xcite . as can be seen in figures [ barufsi]a and [ barufsi]b , the variation of the absolute values of such obtained enhancement factors with the applied potential is severe . + since the physical value of the total cross section can not depend on the off shell features of the potential used in calculations , which are in principal not measurable @xcite , the differences in the magnitude of the @xmath317 factor must reflect itself in a corresponding dependence of the primary production amplitude on the potential used . therefore , it is of crucial importance to realize that the values of the threshold amplitudes @xmath318 are significant only in the context of the potential they were extracted from . figure [ barufsi]c demonstrates , however , that the shapes of the enhancement factors , with the meson exchange mechanism assumed for the primary production , are pretty much the same , independently of the applied @xmath12 potential and correspond to the form of the jost function inferred from the yamaguchi potential . this indicates that the energy dependence of the @xmath12 fsi factors is predominantly determined by the on shell @xmath12 @xmath249 matrix . in references @xcite the formula for the transition matrix element with explicit dependence on the considered off shell features for the initial and final state interaction is derived : @xmath319 \cdot \left [ 1 \,+\ , p_f(\mbox{p},\mbox{k})\right ] \right\ } \ : m_0 \ : \left\{1 \,+\ , \frac{1}{2 } \left[\eta(\mbox{p } ) e^{2i\delta(\mbox{\scriptsize p } ) } \,-\ , 1 \right ] \cdot \left [ 1 \,+\ , p_i(\mbox{p},\mbox{k})\right ] \right\},\ ] ] where subscripts @xmath103 and @xmath320 indicate the initial and final state , respectively . the functions @xmath321 exhibit all the off shell effects of the @xmath12 interaction and the primary production current @xcite and @xmath94 and @xmath11 denote the phase shift and inelasticity , correspondingly . at threshold , the inelasticity in the exit channel is equal to unity ( @xmath322 ) due to the small relative momentum of the outgoing nucleons . the last term of the formula expresses the influence of the initial state interaction on the production process . due to the large relative momenta of the colliding protons needed to create a meson it is characterized by a weak energy dependence in the excess energy range of a few tens of mev . for example in figure [ 1s03p0_eta_q_vps]a one can see that the phase shift variation of the @xmath242 partial wave ( having predominant contribution to the threshold production of pseudoscalar mesons ) in the vicinity of threshold for mesons heavier than @xmath44 is indeed very weak . taking additionally into account that the initial state off shell function @xmath323 is small ( at least it is the case for meson exchange models @xcite ) one reduces the influence of the @xmath12 initial state interaction to the reduction factor @xmath262 which can be estimated from the phase shifts and inelasticities only : @xmath324 at the threshold for @xmath9 meson production this is close to unity since at this energy the inelasticity is still nearly 1 and the @xmath242 phase shift is close to zero ( see figure [ 1s03p0_eta_q_vps]a ) . however , at the @xmath11 threshold , where the phase shift approaches its minimum , the proton proton isi diminishes the total cross section already by a factor of 0.2 @xcite . a similar result was obtained using a meson exchange model for @xmath11 production in the @xmath71 reaction and calculating the proton proton distortion from the coupled channel @xmath325 model @xcite . the authors of reference @xcite concluded that the initial proton proton distortion reduces the total cross section by about a factor of @xmath326 , which keeps constant at least in the studied range of @xmath250 in kinetic beam energy . hence , the closed formula disregarding the off shell effects ( @xmath323 ) permits to estimate the cross section reduction due to the initial state distortion with an accuracy of about @xmath327 . the shape of the @xmath242 phase shift shown in figure [ 1s03p0_eta_q_vps]a indicates that the effect is at most pronounced close to the @xmath11 production threshold , yet for the @xmath158 meson the formula leads to a factor @xmath328 @xcite . the primary production amplitude as well as the off shell effects of the nucleon nucleon fsi ( @xmath329 ) are also weakly energy dependent @xcite , since they account for the short range creation mechanism , which shall be considered in the next section . the accordance of the experimental data with the simple factorization represented by solid lines in figures [ mpppp_cross_pi]b , [ cross_etap_pkl ] and [ cross_eta_fig]a fully confirms the above considerations which imply that the total cross section energy dependence and the occupation density on a dalitz plot is in the first order determined by the on shell scattering of the outgoing particles . however , since the distortion caused by the nucleons is by some orders of magnitude larger than that due to the meson nucleon interaction , even small fractional inaccuracies in the description of nucleon nucleon effects may obscure the inference on the meson nucleon interaction . the differences between the square of the on shell proton proton scattering amplitude and the jost function prescription are presented in figure [ mpppp_cross_pi]a . to minimize the ambiguities that may result from these discrepancies at least for the quantitative estimation of the effects of the unknown meson nucleon interaction one can compare the spectra from the production of one meson to the spectra determined for the production of a meson whose interaction with nucleons is established . for instance , to estimate the strength of the @xmath330 and @xmath331 fsi one can compare the appropriate observables to those of the @xmath332 system . figures [ ratio_and_factors]a and [ ratio_and_factors]b show the dependence of @xmath318 on the phase space volume for @xmath11 and @xmath158 production normalized to @xmath333 . the values of @xmath318 were extracted from the experimental data by means of equation disregarding the proton meson interaction ( @xmath334 was set to 1 ) . if the influence of the neglected interactions were the same in case of @xmath335 and @xmath44 the points would be consistent with the solid line . this is the case for the @xmath255 reaction visualizing the weakness of the proton@xmath158 interaction independently of the prescription used for the proton proton fsi @xcite . in case of the @xmath158 meson energy interaction with the nucleons was expected to be very weak since there exists no baryonic resonance which would decay into @xmath336 @xcite . figure [ ratio_and_factors]a shows independently of the model used for the correction of the proton proton fsi the strong effects of the @xmath330 fsi at low @xmath68 . with the up to date experimental accuracy , from all @xmath337systems the @xmath338 one reveals by far the most interesting features . the dynamics of the @xmath338 system has become recently a subject of theoretical investigations in view of the possible existence of quasi bound or resonant states @xcite . according to reference @xcite within the present inaccuracy of a@xmath339 the existence of quasi bound @xmath11mesic light nuclei could be possible . a direct measure of the formation or non formation of an @xmath11nuclear quasi bound state is the real part of the @xmath11nucleon scattering length @xcite . the determined values of @xmath340 ) range between @xmath341 @xcite and @xmath342 @xcite depending on the analysis method and the reaction studied @xcite , and at present an univocal answer whether the attractive interaction between the @xmath11 meson and nucleons is strong enough to form a quasi bound state is not possible . the shape of the energy dependence of the @xmath343 cross section implies that either the real or imaginary part of the @xmath344 scattering length has to be very large @xcite , which may be associated with a bound @xmath344 system . similarly encouraging are results of reference @xcite , where it is argued that a three body @xmath338 resonant state , which may be formed close to the @xmath345 threshold , may evolve into a quasi bound state for @xmath346 . also the close to threshold enhancement of the total cross section of the @xmath71 reaction was interpreted as being either a borromean ( quasi bound ) or a resonant @xmath330 state @xcite provided that @xmath347 . contrary , recent calculations performed within a three body formalism @xcite indicate that a formation of a three body @xmath338 resonance state is not possible , independently of the @xmath211 scattering parameters . moreover , the authors of reference @xcite exclude the possibility of the existence of an @xmath338 quasi - bound state . results of both calculations @xcite , although performed within a three body formalism , have been based on the assumption of a separability of the two body @xmath211 and @xmath12 interactions . however , in the three body system characterized by the pairwise attractive interactions , the particles can be pulled together so that their two body potentials overlap , which may cause the appearance of qualitatively new features in the @xmath338 system @xcite . specifically interesting is the @xmath345 final state where the pair of nucleons alone is bound by the strong interaction . figure [ ratio_and_factors]d shows the total cross section for the @xmath348 reaction measured close to the production threshold . for excess energies below @xmath73 the data are enhanced over the energy dependence determined for the @xmath124 reaction indicated by the solid curve . this is in qualitative agreement with the calculations of ueda @xcite for the three body @xmath338@xmath325 coupled system , which predict the existence of an @xmath338 quasi bound state with a width of @xmath293 . ueda pointed out that the binding of the @xmath338 system is due to the @xmath53 @xmath211 and @xmath349 @xmath350 interaction which is characterized by no centrifugal repulsion . such repulsion makes the @xmath325 system , in spite of the strong @xmath27 @xmath351 attraction , hard to be bound @xcite . whether the observed cusp at the @xmath114 threshold is large enough to confirm the existence of the @xmath338 bound state is recently vigorously discussed @xcite . the considerations concerning the possible existence of the light @xmath11mesic nuclei will be continued in section [ qbs ] . the enhancement factor for the deuteron@xmath11 interaction inferred from the data in figure [ ratio_and_factors]d varies much stronger in comparison to the proton@xmath11 one , as demonstrated in figure [ ratio_and_factors]c . this suggests that the effects of this interaction should be even more pronounced in the differential distributions of the cross section for the @xmath352 reaction as those observed in case of the @xmath170 process , especially because the `` screening '' from the proton deuteron interaction is by more than an order of magnitude smaller compared to the proton proton interaction as can be deduced from the comparison of the solid lines in figures [ ratio_and_factors]c and [ mpppp_cross_pi]a . the experimental investigations on that issue @xcite as well as the search for the @xmath11mesic nuclei @xcite by measuring proton deuteron induced reactions in the vicinity of the @xmath11 production threshold are on the way . the considerations from the previous section led to the conclusion , that close to the kinematical threshold , the energy dependence of the total cross section is in the first approximation determined via the interaction among the outgoing particles and that the entire production dynamics manifests itself in a single constant , which determines the absolute scale of the total cross section . as a first step towards the understanding of the creation mechanism underlying the production let us compare the total cross sections of neutral pseudoscalar mesons @xmath44 , @xmath11 and @xmath158 . since the masses of these mesons are significantly different , @xmath11 and @xmath158 masses amount to @xmath353 , @xmath354 and @xmath355 , respectively @xcite . ] the influence of the kinematical flux factor f on the total cross section and the suppression due to the initial state interaction @xmath262 depend substantially on the created meson . therefore , for the comparison of the primary dynamics we will correct for these factors and instead of comparing the total cross section we will introduce according to reference @xcite a dimensionless quantity @xmath356 , which depends only on the primary production amplitude @xmath357 and on the final state interaction among the produced particles . close to threshold the initial state interaction , which reduces the total cross section , is dominated by proton proton scattering in the @xmath358 state which may be estimated in terms of phase shifts and inelasticities by employing equation . the @xmath262 factor is close to unity for pion production and amounts to @xmath359 @xcite and @xmath360 @xcite for the @xmath11 and @xmath158 meson , respectively , at threshold . a comparative study of the production of mesons with significantly different masses encounters the difficulty of finding a proper variable at which the observed yield can be compared . the total cross section is defined as the integral over the available phase space volume of transition probabilities reflecting the dynamics of the process from the initial to the final state as written explicitly in equation [ phasespacegeneral ] . thus , if the dynamics of the production process of two different mesons were exactly the same then the above introduced yield would also be strictly the same for both mesons provided it was extracted at the same value of @xmath68 . this inference would however not be valid if the production yields were compared at the @xmath146 or q variables . therefore the volume of the available phase space for the produced particles is the most suited quantity for the regarded comparison @xcite . this could be also the best choice for the investigation of isospin breaking where the cross sections for the production of particles with different masses need to be compared ( e.g. @xmath361 and @xmath362 @xcite ) . figure [ cfisi_vps]a shows the yield of @xmath44 , @xmath11 and @xmath158 mesons in the proton proton interaction as a function of the available phase space volume . the onset of higher partial waves is seen for @xmath44 and @xmath11 mesons in the @xmath68 range between @xmath363 and @xmath364 , whereas the whole range covered by the @xmath158 data seems to be consistent with the pure ss production . one can also recognize that the data for the ss final state are grouped on parallel lines indicating a dependence according to the power law @xmath365 @xcite and that over the relevant range of @xmath68 the dynamics for @xmath158 meson production is about six times weaker than for the @xmath44 meson , which again is a further factor of six weaker than that of the @xmath11 meson . this is an interesting observation , since the quark wave functions of @xmath11 and @xmath158 comprise a similar amount of strangeness ( @xmath366 @xcite ) and hence , in the nucleon nucleon collision one would expect both these mesons to be produced much less copiously than the meson @xmath44 being predominantly built out of _ up _ and _ down _ quarks . on the hadronic level , however , one can qualitatively argue that the @xmath11 meson owes its rich creation in the threshold nucleon nucleon collisions to the existence of the baryonic resonance @xmath367 whose branching ratio into the @xmath368 system amounts to 30@xmath369 @xcite . there is no such established resonance , which may decay into an s wave @xmath370 system @xcite and the @xmath44 meson production with the formation of the intermediate @xmath28 state is strongly suppressed close to threshold , because of conservation laws . prior to the more comprehensive discussion of the production mechanism for different mesons let us present one more interesting observation visualizing that the production dynamics depends strongly not only on the structure of the created meson but also on the isospin configuration of the colliding nucleons . for example , in the case of the @xmath11 meson the comparison of the cross section from the production in proton proton and proton neutron collisions revealed that the yield of the @xmath11 meson in the isosinglet nucleon configuration exceeds the one in the isospin triplet state by more than an order of magnitude . the ratio @xmath371 presented in figure [ cfisi_vps]b , was determined to be about 6.5 in the excess energy range between @xmath372 and @xmath373 @xcite . this implies that the production of the @xmath11 meson with the total isospin equal to zero exceeds the production with isospin @xmath374 by a factor of 12 , since @xmath375 and @xmath376 and hence @xmath377 . on the mesonic level , this large difference of the total cross section between the isospin channels suggests the dominance of isovector meson ( @xmath9 and @xmath378 ) exchanges ( figure [ graph_mesonexchange ] graph b ) in the creation of @xmath11 in nucleon nucleon collisions @xcite , but alternatively it can also be explained on the quark gluon level assuming the instanton induced flavour dependent quark gluon interaction @xcite ( see below figure [ dillig_gluons ] graph e ) . the above examples visualize one of the most interesting problems in the investigations of the dynamics of the close to threshold meson production : namely the determination of the relevant degrees of freedom for the description of the nucleon nucleon interaction , especially in case when the nucleons are very close together . as pointed out by nakayama @xcite , the transition region from the hadronic to constituent quark degrees of freedom does not have a well defined boundary and at present both approaches should be evaluated in order to test their relevance in the description of close to threshold meson production in the collision of nucleons . simple geometrical considerations presented in figure [ cartoon ] indicate that at distances smaller than @xmath379 the internucleon potential should begin to be free of meson exchange effects and may be dominated by the residual colour forces @xcite . in the previous section it was shown , that the close to threshold production of mesons occurs when the colliding nucleons approach distances of about @xmath380 in case of @xmath44 and of about @xmath47 in case of @xmath45 production ( see table [ momtranstable ] ) . this distance is by about an order of magnitude smaller than @xmath379 and it is rather difficult to imagine in coordinate space an exchange of mesons between nucleons as mechanism of the creation process . such small collision parameters imply that the interacting nucleons objects of about 1 fm overlap and their internal degrees of freedom may be of importance . nonetheless , in the last decade , the effective theory based on the meson exchanges , which accounts for the size of the participating particles by introduction of the momentum transfer dependent form factors , has been extensively utilized for the description of the creation process . figure [ graph_mesonexchange ] represents the regarded mechanisms . the question which processes on the mesonic level are responsible for the @xmath44 production was considered already in 1966 by koltun and reitan @xcite . the authors deemed the direct pion production on one of the protons ( fig . [ graph_mesonexchange]a ) to play the most important role in the @xmath381 reaction . this supposition was anticipated until the precise experiments performed at iucf @xcite and celsius @xcite revealed that this mechanism can not account for more than @xmath382 of the total yield , as demonstrated in figure [ haidenbauer_acta ] . subsequent investigations showed that the inclusion of the processes illustrated in figure [ graph_mesonexchange]b , when the pion is produced on one of the protons and then scatters on the other one ( in the @xmath351 s wave @xcite and @xmath351 p wave via @xmath28 @xcite ) , increases the cross section essentially yet not enough to describe the data , even if the off shell properties of the @xmath351 amplitude are taken into account @xcite . the contribution of the rescattering mechanism is shown as the dashed line in figure [ haidenbauer_acta ] . the agreement with the data can , however , be achieved as shown by the solid line if one takes additionally into account the mechanisms of heavy meson exchange associated with an intermediate virtual nucleon antinucleon state @xcite illustrated in figure [ graph_mesonexchange]c . a thorough discussion concerning @xmath9 meson production in differently charged channels can be found e.g. in references @xcite . the dominance of the heavy meson exchange and the off shell @xmath351 rescattering mechanisms , both reflecting the short range nucleon nucleon interactions , confirms the simple estimation of the reaction region based on the momentum transfer between the colliding nucleons . thus , in case of the @xmath44 meson creation , this approach succeeded to describe the magnitude of the total cross section . however , the meson - exchange model at the present stage does not reproduce the angular distributions of the spin observables for all @xmath383 isospin channels , simultaneously . specifically , though the only available calculations of the spin correlation coefficients based on the jlich model @xcite are in a good agreement with the experimental data for the reactions @xmath384 @xcite and @xmath385 @xcite they reveal serious discrepancies when confronted with the angular dependences of the spin correlation coefficients and @xmath386 of the @xmath159 reaction @xcite . therefore , further investigations concnerning the relative contributions of various mechanisms , as well as the different partial waves are still required . + the role of the mechanisms leading to the @xmath9 production have been also considered in context of the lowest order chiral perturbation theory @xcite and relativistic impulse approximation @xcite . in contrast to the meson exchange models the chiral perturbation method results @xcite in the destructive interference between the direct and the pion - rescattering terms , and consequently implies more significant contributions originating from the shorter - than - pion - range mechanisms , like the heavy - meson exchanges or the creation of pions in the fusion of exchanged mesons @xcite . in case of the @xmath11 and @xmath158 mesons the investigations of the mechanisms underlying the production process are even more difficult . it is partly due to the fact that in contrary to the pion case even the coupling constants determining the strength of the @xmath387 vertex are poorly established . the values derived by various methods differ by more than an order of magnitude @xcite . for example , the estimation based on the dispersion relation results in @xmath388 smaller than 1 @xcite , whereas the value obtained from the fits to low energy nucleon nucleon scattering in one boson exchange models amounts to 7.3 @xcite . this uncertainty prevents from the exact estimation of the contribution from the nucleonic current . however , according to the systematic investigations of nakayama et al . @xcite this mechanism is of minor importance . the calculations have been performed with @xmath389 used in the @xmath12 scattering analysis @xcite which is close to the upper limit of the predicted values . + it is at present rather well established @xcite that the @xmath11 meson is produced predominantly via the excitation of the @xmath53 baryonic resonance @xmath367 which subsequently decays into @xmath11 and nucleon and whose creation is induced through the exchange of @xmath9 , @xmath11 , @xmath378 , @xmath26 and @xmath390 mesons , as shown in figure [ graph_mesonexchange]e . though all the quoted groups reproduce the magnitude of the total cross section , their models differ significantly as far as the relative contribution from the @xmath9 , @xmath11 and @xmath378 exchange mechanisms is concerned . the discrepancies are due to the not well known strength of @xmath79@xmath37@xmath53 couplings and the @xmath211 scattering potential . for example , while the dominance of the @xmath378 meson exchange is anticipated by authors of references @xcite , it is rather the exchange of the @xmath11 meson which dominates the production if one takes into account effects of the off shell @xmath211 scattering @xcite , or utilizes the multi channel multi resonance model @xcite . in any case , the hitherto performed studies aiming to describe the total cross section show that the close to threshold production of @xmath11 mesons in nucleon nucleon collisons is dominated by the intermediate virtual @xmath53 nucleon isobar whose width overlaps with the threshold . in order to disentangle between various scenarios of the @xmath53 excitation a confrontation of the predictions with other observables is needed . the interference between considered amplitudes causes a different behaviour depending on the assumed scenario e.g. of the @xmath11 meson angular distributions . these differences , however , are too weak in the close to threshold region to judge between different models . also the ratio of the @xmath11 meson production via the reactions @xmath391 and @xmath392 can be equally well described either by assuming the @xmath378 meson exchange dominance @xcite or by taking pseudoscalar and vector mesons for exciting the @xmath53 resonance @xcite . in the latter case , shown in figure [ nakayama_eta ] , the excitation of the resonance via the @xmath378 meson exchange was found to be negligible . yet , the predictions of the analyzing power depends crucially on the assumed mechanism @xcite . this fact has triggered already experimental investigations which aim to determine the spin observables @xcite . figure [ nakayama_eta ] presents also a contribution from the mesonic current where the @xmath11 meson is created in the fusion of virtual e.g. @xmath378 or @xmath390 mesons emitted from both colliding nucleons . in comparison to the overwhelming strength of the resonance current this mechanism is however by a factor of 30 too weak . in contrast , it is enough to explain the magnitude of the close to threshold @xmath158 meson production in the proton proton interaction , which is just by about a factor of 30 smaller compared to the @xmath11 meson . among the mesonic currents regarded by authors of reference @xcite the @xmath393 gives the dominant contribution , by a factor of five larger than the @xmath394 and @xmath395exchange . however , the understanding of the @xmath158 production on the hadronic level is far from being satisfactory . the magnitude of the total cross section was also well reproduced in the frame of a one boson exchange model ( nucleonic current ) where the virtual boson ( @xmath396 ) created on one of the colliding protons converts to the @xmath158 on the other one @xcite . taking into account the off shell effects of the @xmath397 amplitude it was found that the short range @xmath26 and @xmath378 meson exchanges dominate the creation process , whereas the @xmath9 exchange plays a minor role . on the contrary , other authors @xcite reproduced the magnitude of the total cross section with @xmath9 exchange only and concluded that the @xmath11 , @xmath378 and @xmath390 exchange currents either play no role or cancel each other . however , in both of the above quoted calculations @xcite the initial state interaction between protons , which reduces the rate by a factor of about 3 , was not taken into account and hence the obtained results would in any case reproduce only @xmath398 of the entire magnitude and could be at least qualitatively reconciled with the mentioned result of reference @xcite , where the nucleonic current was found to be small . moreover , the choice of the another prescription for the form factors could reduce the one pion exchange contribution substantially @xcite . however , the picture that the @xmath158 meson is predominantly created through the mesonic current , which we would like to advocate , remains at present unclear as well . this is because the magnitude of the total cross section could have been also described assuming that the production of @xmath158 is resonant @xcite . as possible intermediary resonances the recently reported @xcite @xmath399 and @xmath400 have been considered . these resonances were deduced from @xmath158 photoproduction data , under the assumption that the close to threshold enhancement observed for the @xmath401 reaction can be utterly assigned to resonance production . further , as well strong , assumptions have been made in the derivation of the @xmath402 and @xmath403 coupling constants @xcite . hence , it is rather fair to state that in the case of the close to threshold @xmath44 and @xmath11 meson production in nucleon nucleon collisons the dynamics is roughly understood on the hadronic level , but the mechanisms leading to the @xmath158 creation are still relatively unknown . unil now it was not possible to estimate satisfactorily the relative contributions of the nucleonic , mesonic and resonance current to the production process . in fact , model uncertainties allow that each one separately could describe the absolute values of the @xmath255 total cross section . this rather pessimistic conclusion calls for further experimental and theoretical research . the understanding of the production dynamics of the @xmath158 meson on both hadronic and the quark gluon level is particularly important since its wave function comprises a significant gluonic component @xcite , distinguishing it from other mesons and hence the comprehension of the mechanism leading to its creation in collisions of hadrons may help in the determination of its quark gluon structure which is also of importance to investigate possible glueball candidates @xcite . hereafter we will briefly report about the gluonic mechanisms which may additionally to the meson exchange processes discussed above contribute to the @xmath158 production . as already mentioned , the close to threshold production of @xmath11 and @xmath158 mesons in the nucleon nucleon interaction requires a large momentum transfer between the nucleons and hence can occur only at distances of about @xmath404 ( see table [ momtranstable ] ) . this suggests that the quark gluon degrees of freedom may indeed play a significant role in the production dynamics of these mesons . a possibly large glue content of the @xmath158 and the dominant flavour singlet combination of its quark wave function may cause that the dynamics of its production process in nucleon nucleon collisions is significantly different from that responsible for the production of other mesons . in particular , the @xmath158 meson can be efficiently created via a `` contact interaction '' from the glue which is excited in the interaction region of the colliding nucleons @xcite . a gluon induced contact interaction contributing to the close to threshold @xmath255 reaction derived in the frame of the u(1)anomaly extended chiral lagrangian is discussed in references @xcite . the strength of this contact term is related to the amount of spin carried by polarized gluons in a polarized proton @xcite , thus making the study of the close to threshold @xmath158 meson production even more interesting . figure [ graf_glue ] depicts possible short range mechanisms which may lead to the creation of the @xmath158 meson via a fusion of gluons emitted from the exchanged quarks of the colliding protons @xcite or via an exchange of a colour singlet object made up from glue , which then re scatters and converts into @xmath158 @xcite . the hadronization of gluons to the @xmath158 meson may proceed directly via its gluonic component or through its overwhelming flavour singlet admixture @xmath405 ( see fig . [ triangle ] ) . contrary to the significant meson exchange mechanisms and the fusion of gluons of figure [ graf_glue ] graph a ) , the creation through the colour singlet object proposed by s.d . bass ( graph [ graf_glue]b ) is isospin independent , and hence should lead to the same production yield of the @xmath158 meson in both reactions ( @xmath255 and @xmath406 ) because gluons do not distinguish between flavours . this property should allow to test the relevance of a short range gluonic term @xcite by the experimental determination of the cross section ratio @xmath407 , which in that case should be close to unity after correcting for the final and initial state interaction between participating baryons . the other extreme scenario assuming the dominance of the isovector meson exchange mechanism should result in the value of @xmath408 close to 6.5 as was already established in the case of the @xmath11 meson ( see fig . [ cfisi_vps]b ) . the experimental investigations of that issue are in preparation @xcite . in addition to the qcd motivated investigation presented above @xcite , the interesting features of the close to threshold meson production , in particular the large cross section for the @xmath11 meson in proton proton interactions exceeding the one of the pion , or even more surprisingly large cross section of the @xmath11 production in proton neutron collisions , encouraged also other authors to seek for the underlying ozi rule violating creation mechanisms in the frame of microscopic models of qcd @xcite . the hitherto regarded processes are presented in figure [ dillig_gluons ] . as indicated in the pictures the structure of participating baryons has been modeled as quark diquark objects with harmonic confinement @xcite . in the upper graphs the large momentum transfer is shared by the exchanged gluons with a subsequent interchange of quarks to provide a colourless object in the final state . the lower graphs depict the two examples of instanton induced interactions with a 6quark antiquark ( fig . [ dillig_gluons]d ) and two - gluon vertex ( fig . [ dillig_gluons]e ) . adjusting the normalization to the cross section of the @xmath234 reaction at a single energy point the model @xcite accounts roughly for close to threshold cross sections of other pseudoscalar and vector mesons in proton proton collisions . though this approach is characterized by significantly smaller number of parameters than in the meson exchange models , their uncertainties allow for the description of the data equally well either by the gluon exchange or by the instanton induced interactions , at least in case of the @xmath44 , @xmath158 , @xmath390 and @xmath45 mesons . yet for the resonance dominated @xmath11 and @xmath284 meson production in proton proton collisions it was found that the instanton induced interaction presented by graph [ dillig_gluons]d is not sufficient . similarly , the authors of reference @xcite argue that the instanton induced interaction with a quark gluon vertex ( graph [ dillig_gluons]e ) should be of no importance for the @xmath11 production in proton proton collisions . this arises from the properties of the vertex which lead to the @xmath11 production only in case of an interaction between quarks of different flavours ( @xmath409 ) and in the quark diquark model of baryons the proton consists of @xmath410diquark and @xmath411quark and correspondingly the neutron is modeled as @xmath410diquark and @xmath56quark . although negligible in case of proton proton interactions , this mechanism may contribute significantly to the total cross section of the @xmath412 reaction and indeed as shown in reference @xcite it reproduces the data . however , the magnitude of the cross section is very sensitive to the size of the instanton in the qcd vacuum and similar to the uncertainty of coupling constants in case of the hadronic approach at present it makes precise predictions impossible . thus , despite the partial successes , at the present stage of developments both approaches on the quark gluon and on the hadronic level do not deliver an unambiguous answer to the dynamics of the close to threshold meson production in the nucleon nucleon interaction . vector meson production studies in the mass range up to @xmath413 in hadronic collisions are especially motivated by questions concerning the strangeness content of the nucleon . within the quark model the @xmath45 meson is an ( almost ) ideally mixed @xmath414 state , whereas the nucleon is composed only of @xmath411 and @xmath56quarks . according to the ozi rule @xcite , @xmath45 production in the nucleon nucleon interaction would be forbidden as a process with disconnected quark lines , if the @xmath45 was a pure @xmath414 state . however , due to a small deviation from ideal mixing ( @xmath415 @xcite ) the @xmath45 meson is allowed to couple to the nucleon through its @xmath416 admixture . quantitatively , the naive ozi rule results in a suppression of @xmath45 compared to @xmath390 production in hadronic interactions @xmath417 after phase space corrections , where @xmath418 , @xmath38 and @xmath13 denote hadronic systems consisting only of light quarks @xcite . any significant deviation might be interpreted as a hint for a strangeness component in the nucleon . a considerable contribution of strange sea quarks to the nucleon wave function is suggested by results on the @xmath419 term in pion nucleon elastic scattering @xcite , on the nucleon s structure function in deep inelastic scattering with polarized muons @xcite and , more recently , on charm production in deep inelastic neutrino scattering @xcite . large violations of the ozi expectation ( eq . ) have been found in @xmath420 annihilation experiments by the asterix , crystal barrel and obelix collaborations at lear ( as reviewed in @xcite ) . however , deviations are predominantly found in @xmath420 annihilation at rest as compared to higher energies and seem to be restricted to s wave processes only . furthermore , the results are strongly dependent on the final state : while the @xmath421 and @xmath422 channels exceed the ozi value by a factor of 20 and 100 , respectively , little or no effect is seen in the @xmath423 and @xmath424 final states . these data have been interpreted as `` shake out '' and `` re - arrangement '' of a negatively polarized @xmath414 fock space component of the proton wave function @xcite , which also accounts for large double @xmath425 cross sections @xcite in @xmath420 annihilation measured at the jetset experiment @xcite . however , ozi allowed two step processes via intermediate @xmath426 or @xmath427 states have been shown to describe available data from @xmath420 annihilation without any strangeness in the nucleon @xcite . in nucleon nucleon scattering , effects of competing two step processes for @xmath45 production via intermediate @xmath428 or @xmath429 states are expected to be of minor importance @xcite . furthermore , in close to threshold @xmath45 production the entrance channel must be in a @xmath430 state ( see table [ partialtransitions ] ) due to parity and angular momentum conservation and it is the spin triplet fraction in @xmath420 annihilation which is strongly correlated to the @xmath45 meson yield @xcite . thus , cross section ratios for @xmath45 and @xmath390 production in proton proton scattering should clearly indicate possible ozi violations and probe the @xmath414 component of the nucleon . the exclusive production ratio has been determined in the reaction @xmath431 at a beam momentum of @xmath432 corresponding to excess energies of @xmath433 and @xmath434 to the @xmath45 and @xmath390 thresholds , respectively , by the disto collaboration at saturne @xcite . after phase space corrections , the observed ratio exceeds the ozi expectation by an order of magnitude , while data at excess energies larger than @xmath435 @xcite show an enhancement smaller by a factor of four ( see fig . [ phiomega ] ) . one boson exchange models @xcite underestimate the absolute value of the @xmath45 production cross section measured at the disto experiment by at least a factor of three . + figure [ phiomega ] includes results obtained within a one pion exchange model @xcite , which provides a reasonable description of high energy data . model predictions taking into account the proton proton final state interaction only falls short of the recent close to threshold value @xcite . using instead a parametrization of the energy dependence of the @xmath390 production cross section and considering both the finite @xmath390 width as well as effects of proton proton fsi @xcite , the @xmath436 ratio can be determined at equal excess energies , thus reducing uncertainties due to the available phase space , partial wave amplitudes angular distribution relative to the @xmath50 system measured at an excess energy of @xmath434 shows evidence for p and d waves contributing @xcite . ] and effects of the proton proton fsi . in conclusion , the excess over the ozi prediction is decreased to a factor of five @xcite , i.e. to a modest enhancement as compared to the results from @xmath420 annihilation experiments @xcite and the recent cosy tof result at an excess energy of @xmath437 @xcite leads to a similar enhancement of @xmath438 compared to the ozi expectation . ] . in the disto results , @xmath45 production at an excess energy of @xmath433 appears to be dominated by s wave relative to the nucleon nucleon system in agreement with theoretical predictions @xcite and phenomenologically described in section [ mwhs ] . however , the proton proton angular distribution relative to the @xmath45 direction is found to deviate significantly from isotropic emission at this beam energy with evidence for a p wave contribution in the proton proton subsystem . consequently , @xmath45 production at @xmath433 above threshold , although dominated by a @xmath430 proton proton entrance channel , partly proceeds via the @xmath16 and @xmath439 spin singlet partial waves @xcite . thus , the influence of the spin triplet entrance channel , prerequisite to a `` re arrangement '' of an @xmath414 component in the nucleon wave function @xcite , might be diluted in the presently available data and would be expected to be more prominent in data closer to the production threshold . recently , within the framework of a relativistic meson exchange model , nucleonic and meson exchange currents have been used explicitly to study vector meson production in nucleon nucleon scattering ( fig . [ kanzo_curr ] @xcite ) . uncertainties in the model calculation for the nucleonic current arise mainly from a large range of values quoted for vector and tensor @xmath440 coupling constants @xcite , as well as from the vector and tensor vertex form factors for the coupling of an on mass shell vector meson to off mass shell nucleons . note , that a resonance contribution to the nucleonic current ( fig . [ kanzo_curr]a ) is neglected in the model @xcite , since there are no isospin1/2 nucleon resonances experimentally known coupling to the @xmath441 channel @xcite . due to the strength of both the @xmath442 and @xmath443 couplings and the @xmath444 coupling constant as compared to other possible choices , only the @xmath445 vertex is considered in evaluating the meson exchange current . for @xmath390 production , contributions from the latter and the nucleonic current are found to interfere destructively when added coherently , which leaves two choices for a given value of the @xmath446 tensor to vector coupling ratio when fitting the close to threshold data on the reaction @xmath447 taken at saturne @xcite for the only remaining free parameter of the model , the cut off parameter at the @xmath446 form factor . effects from a destructive interference of mesonic exchange current and nucleonic current contribution are also concluded by other authors for @xmath390 production in @xcite ) . the dominant nucleonic current contribution arises from @xmath448 exchange . however , the combined result of the studied amplitudes shows similarly to @xcite a dependence on the choice of the tensor to vector coupling ratio . ] and for @xmath45 production in @xcite . in nakayama s analysis the energy dependence of the total cross section is slightly influenced by either choice , as shown by the solid and dashed curves @xcite , with cut off masses of @xmath449 ( dashed line ) and @xmath450 ( solid line ) , respectively . ] in figure [ omega_cross ] together with a parametrization based on one pion exchange ( dash dotted line @xcite ) . in comparison , total cross section data taken at the spes 3 spectrometer up to excess energies of @xmath451 @xcite , the disto data point discussed above @xcite , high energy data from @xcite and recent results obtained at the tof facility at cosy at excess energies of @xmath452 and @xmath453 , respectively @xcite , are included . the one pion exchange parametrization @xcite deviates strongly from the close to threshold data , where fsi effects become important , which have not been included in the calculation . both predictions obtained within the relativistic meson exchange model of @xcite overestimate the recent cosy tof data @xcite by at least a factor of five . the solid curve in figure [ omega_cross ] seems to follow the data slightly better , which would correspond to a lower cut off parameter in the nucleonic current and , consequently , a dominating meson exchange current contribution . within the model @xcite , angular distributions of the produced mesons are expected to differ depending on the dominant production mechanism for both @xmath390 and @xmath45 production : while the mesonic current leads to an isotropic angular distribution , the nucleonic current results in a @xmath454 behaviour for the vector meson emission angle @xmath455 in the overall centre of mass system behaviour arises according to @xcite from the spin dependent part of the nucleonic current contribution , the `` magnetization current '' @xcite . ] . thus , angular distributions offer a unique possibility to separate both contributions , or to isolate the nucleonic current contribution in order to determine the @xmath456 coupling directly in a combined analysis of @xmath390 and @xmath45 production data in @xmath50 induced reactions : the experimentally observed isotropic @xmath45 angular distribution @xcite clearly favours a dominant @xmath457 meson exchange over the nucleonic current @xcite . contrary , the centre of mass angular distribution of the @xmath390 meson obtained at the cosy tof facility at an excess energy of @xmath453 @xcite exhibits a strong anisotropy similar to the observation by the disto collaboration at an excess energy of @xmath434 @xcite , indicating the presence of higher partial waves and rather favours a dominant nucleonic current contribution instead of meson exchange currents suggested by the shape of the energy dependence of the total cross section in figure [ omega_cross ] . further constraints on the relevant vector meson production mechanisms in nucleon nucleon scattering might be obtained from comparing with complementary meson induced reactions : an analysis of @xmath458 data concludes @xmath390 production to result from strong interferences between the meson exchange current and both nucleonic and resonance currents , which contribute with similar strengths @xcite . near threshold , the @xmath24 @xmath25 resonance and the @xmath53 @xmath367 and @xmath459 excitations are found to contribute , with the @xmath53 contributions canceling to a large extent couplings and relative phases have been taken from @xcite . ] . in conclusion , at least for the reaction @xmath460 , the nucleonic current contribution appears to be much larger for @xmath390 production relative to the meson exchange current , contrary to @xmath45 production @xcite . thus , the scenario from pion induced vector meson production seems to agree qualitatively with the observations in nucleon nucleon scattering , and is interesting to note especially in view of the @xmath390 angular distribution from the cosy tof data @xcite . however , angular distributions at lower excess energies would clearly help to constrain the interpretation of the presently available nucleon nucleon data in future . in a combined analysis of @xmath390 and @xmath45 production within the relativistic meson exchange framework @xcite a value of the @xmath45 coupling to the nucleon has been extracted from the disto data and is found to be consistent with the ozi expectation @xmath461 @xcite . yet , as stated by nakayama , the value still has large uncertainties , especially due to the lack of more and precise data in the threshold region @xcite . indeed , on the basis of extracted transition matrix elements @xcite combining @xmath351 and @xmath462 induced reactions , a @xmath436 production ratio has been derived , which exceeds the ozi estimate of eq . by a factor of five in both cases . as suggested in @xcite , future experiments on @xmath45 and @xmath390 production in close to threshold proton neutron collisions might provide further evidence for the intrinsic strangeness content of the nucleon , as predictions for @xmath45 production cross sections based on the `` shake out '' and `` rearrangement '' mechanism @xcite and on meson exchange models @xcite differ by more than an order of magnitude . the @xmath413 meson mass range is continuously under discussion regarding the nature of the scalar resonances @xmath463 and @xmath464 , which have been interpreted as exotic two quark two antiquark states @xcite , possible glueball candidates @xcite , excitations of the light quark vacuum condensate ( _ minions _ @xcite ) , conventional @xmath465 states @xcite , or molecular like @xmath426 bound states @xcite . just already the similarity of the masses of these objects with twice the kaon mass , see figure [ mass_pdb_kk_interaction]a , suggests a strong correlation between these resonances and the @xmath10@xmath466 system . within the framework of the jlich meson exchange model for @xmath467 @xcite and @xmath468 @xcite scattering the @xmath426 interaction dominated by vector meson exchange gives rise to a @xmath426 bound state identified with the @xmath463 in the isoscalar sector , while the isovector @xmath464 is concluded to be a dynamically generated threshold effect @xcite . as shown in figure [ mass_pdb_kk_interaction]b , both shape and absolute scale of @xmath469 transitions turn out to depend crucially on the strength of the @xmath426 interaction , which , in turn is prerequisite of a @xmath426 bound state interpretation of the @xmath463 . similar effects might be expected for the elementary kaon antikaon production in proton proton scattering . although the @xmath470 decay mode of @xmath463 and @xmath464 is rather weak in comparison to the dominant @xmath471 and @xmath472 decay channels @xcite , even a new theoretical analysis based on the chiral approach can not account for the @xmath463 and @xmath464 if the @xmath426 channel is not introduced additionally to the @xmath471 and @xmath472 interaction @xcite . an analysis of the @xmath471 and @xmath470 interaction @xcite showed that the @xmath473 corresponds to two poles on the second and third riemann sheet , respectively , and appears physically as an object with a decay width of about @xmath474 and a narrow peak width of about @xmath475 . the same parameters of the @xmath473 were found by utilizing a unitarized quark model , according to which the @xmath473 was interpreted as a @xmath476 state with a large admixture of a @xmath470 virtual state @xcite . the origin of the scalar resonances was also thoroughly studied by means of a coupled channel analysis considering @xmath471 , @xmath470 and @xmath477 meson meson scattering @xcite . decreasing gradually the interchannel coupling constants it was inferred that for some solutions at the limit of the fully uncoupled case the @xmath473 corresponds to a @xmath470 bound state @xcite . recently notable progress has been made concerning the discussion of this topic due to approaches in unitary extensions of chiral perturbation theory where the resonances under consideration qualify as dynamically generated resonances from the multiple scattering with the lowest order chiral lagrangian , as outlined in the review @xcite . in high energy experiments , the @xmath473 meson is observed as a resonance in the system of two pions produced in a variety of hadro production reactions @xcite and in the hadronic decays of either heavier mesons @xcite or the @xmath478 boson @xcite created in @xmath479 collisions . these experiments study the invariant masses of the created neutral ( @xmath480 ) @xcite and charged ( @xmath481 ) @xcite pion pairs . similarly , charged @xcite and neutral @xmath482 @xcite mesons were observed as a clear signal in an invariant mass spectrum of the @xmath483 system . complementary to these approaches , studying the interaction of @xmath471 , @xmath470 and @xmath472 meson pairs , the cosy11 collaboration presently investigates the possible manifestation of the mesons @xmath473 and/or @xmath482 as doorway states leading to meson production in proton proton collisions , namely @xmath484 . by measuring the missing mass of the @xmath50system the overlapping @xmath473@xmath482 resonances are studied as a genuine particle produced directly at the reaction . the production of these resonances is carried out at a mass range a few tens of mev below the @xmath470 threshold ( where the resonances can only decay into non strange mesons ) and above the @xmath470 thresholds , where the equivalent part of these broad resonances can be excited . it is obvious that for the production of a broad resonance the phrase `` close to threshold '' is not well defined and implies here that the beam momentum is such that masses just in the range of the resonance can be excited . different mass ranges of such objects are produced with different excess energies making the interpretation of the experimental data difficult but still possible @xcite . it is worth noting , that recently first measurements of the @xmath473 meson production relatively close to its threshold but still above the @xmath470 threshold were performed by observing the @xmath45 meson decays into the @xmath485 channel @xcite , where the @xmath45 is only about @xmath486 heavier than the @xmath473 . in a recent publication @xcite data on the close to threshold @xmath487 production following the proton proton interaction at an excess energy of @xmath488 are presented . the obtained distribution of the missing mass to the @xmath50 system is shown in figure [ mm_kk_f0 ] and demonstrates that the non resonant @xmath487 production ( solid line ) is hardly distinguishable from the resonant @xmath489 reaction sequence ( dashed line ) @xcite . since the statistics of the data were not sufficient to favour one of the two processes , the cross section was extracted for both and resulted in nearly identical values of : @xmath490 and @xmath491 including statistical and systematical errors , respectively . the issue , whether there is a chance to distinguish between @xmath470 pairs originating from the decay of a genuine @xmath473/@xmath482 resonance and those produced by a strong @xmath471@xmath470 correlation is at present under theoretical investigation @xcite . still , calculations based upon the one pion exchange and a breit wigner presentation of the @xmath473 resonance indicate @xcite that no @xmath473 signal might be extracted from the @xmath487 invariant mass spectrum in @xmath492 reaction at near threshold energies , due to the large contribution from other reaction channels @xcite as arising from pion and kaon exchange . considering the @xmath493 as a conventional @xmath494 meson , predictions on the total and differential cross sections for @xmath482 production via the reaction @xmath495 are given for the cosy energy range in @xcite : within a two step model , where the @xmath482 is found to be produced predominantly from u channel @xmath9 exchange with a subsequent fusion of the nucleons to a deuteron bound state , a rather large total cross section is concluded . an experimental verification at the anke facility @xcite might give an additional hint on the @xmath494 content of the isovector scalar resonance . the measurements also aim at determining the rather poorly known @xmath482 branching ratios to the @xmath470 and @xmath483 final states ( see discussion in @xcite and references therein ) . the role of final state interactions in the reactions @xmath496 and @xmath497 is explicitly studied in @xcite : considering primary production amplitudes within meson baryon chiral perturbation theory , @xmath498 final state interactions due to the proximity of the @xmath499 to the @xmath500 system are expected to change the cross section by up to an order of magnitude and to significantly influence invariant mass distributions . the @xmath501 and @xmath502 final state systems are investigated within a coupled channel chiral unitary approach , in which the @xmath493 is generated dynamically as a meson meson resonance . once the two remaining independent model parameters related to the couplings of the two pseudoscalar meson systems are fixed by sensitive experimental data , i.e. @xmath501 and @xmath498 invariant mass distributions , the approach will provide absolute predictions for the @xmath503 system . thus , the underlying dynamics of the coupled channel system giving rise to the @xmath493 should be probed directly by the experimental approach of @xcite . further experimental efforts on studying the iso singlet @xmath473 and the iso triplet @xmath482 scalar resonances are continuing at the cosy11 @xcite and anke @xcite facilities at cosy using different approaches . in elementary hadronic interactions with no strange valence quark in the initial state the associated strangeness production provides a powerful tool to study reaction dynamics by introducing an `` impurity '' to hadronic matter . thus , quark model descriptions might be related to mesonic or baryonic degrees of freedom , with the onset of quark degrees of freedom expected for kinematical situations with large enough transverse momentum transfer . a large fraction of the nucleon spin is carried by the intrinsic orbital angular momentum of quarks which in the condensate are pairwise in a @xmath504 state . the challenge of the strange quark , with its intermediate mass between light and heavy quarks , is to explore the effective degrees of freedom via the associated strangeness production in the low energy structure of qcd . the @xmath505 ( @xmath506 ) vacuum condensate is the link between current quark and gluon of qcd and valence quark and potentials of the quark model , structures which can give the massless qcd quarks their constituent masses . in close to threshold kaon hyperon production , effects of the low energy hyperon nucleon interaction are inherent to the observables , and allow to constrain hyperon nucleon interaction models . as the strange quark is heavy compared to up and down quarks ( @xmath507 ) it is not a priori clear , whether the strange quark is appropriately treated as a light quark on the hadronic scale in chiral perturbative approaches with an expansion parameter of @xmath508 ( with the kaon mass @xmath509 and the pion decay constant @xmath510 ) compared to @xmath511 . in exploratory calculations , su(3 ) baryon chiral perturbation theory only has been applied to kaon photo and electroproduction @xcite , for a more detailed discussion see @xcite ) . with a significant contribution of strange sea quarks to the nucleon wave function being suggested by complementary experimental approaches during the past years ( see section [ vmp ] and @xcite ) the creation of an @xmath414 pair in the nucleon nucleon interaction should provide a further piece of valuable information on the structure of the nucleon @xcite . especially polarization observables are expected to act as very sensitive tools @xcite . at the same time , both the kaon and antikaon ( see section [ sdikkp ] ) production cross section in nucleon nucleon scattering are important input parameters within the framework of transport model calculations used to describe the strangeness flow in heavy ion collisions : due to strangeness conservation , kaons will not be absorbed in the nuclear medium . thus , kaon yields in subthreshold production are expected to probe the early stage of the collision , i.e. hot , dense matter , the nuclear equation of state at high densities @xcite and the in medium kaon potential @xcite , see also section [ sdikkp ] ) , which relates to the partial restoration of chiral symmetry @xcite ( for a review see @xcite ) . elementary cross sections are also of interest in view of the recent inclusive subthreshold measurements in nucleus nucleus collisions , resulting in comparable @xmath284 and @xmath83 yields at the same energy per nucleon below threshold @xcite . exclusive data close to threshold on the elementary kaon production via @xmath512 have only become available recently with the advent of dedicated experiments at cosy . figure [ pkl_cross ] shows the available total cross section data obtained at the cosy11 @xcite and cosy tof facility @xcite together with an earlier result from the bnl cosmotron @xcite and theoretical calculations and parametrizations as a function of the excess energy q. + assuming the production amplitude to be constant and both the @xmath513 system and the @xmath284 relative to the @xmath513 subsystem to be in s wave leads to a production cross section proportional to @xmath514 ( section [ siv ] , eq . ) . this dependence has been adopted by schrmann and zwermann @xcite for a quartic ( in terms of the maximum meson momentum ) parametrization of the cross section ( black solid line in fig . [ pkl_cross ] ) , leading to a reasonable fit @xcite of data at excess energies above @xmath515 . although quite commonly employed in the description of nucleon nucleus and nucleus nucleus collisions in the past @xcite , the quartic parametrization underestimates the close to threshold data by an order of magnitude , as does a linear parametrization once suggested by randrup and ko ( grey solid line @xcite ) , overestimating the presently available data at small excess energies by more than two orders of magnitude . it should be noted that both descriptions have been estimated from high energy data , before any of the near threshold measurements were done . close to threshold , deviations from the @xmath514 energy dependence result from the strong s wave final state interactions in the three two body @xmath516 , @xmath513 and @xmath517 subsystems . as the strong interaction in the @xmath516 and @xmath517 systems appears to be weaker by an order of magnitude as compared to the proton@xmath518 case @xcite , an accurate description of the energy dependence is expected to be achieved by considering the strong attractive @xmath513fsi and repulsive coulomb corrections in the proton kaon subsystem . following the discussion in section [ mwhs ] , a reasonable fit of the energy dependence is obtained by factorizing the reaction amplitude @xmath519 according to eq . : @xmath520 here , @xmath357 is the short range production amplitude , which is assumed to be constant in the energy range close to threshold , @xmath261 and @xmath262 describe the elastic interaction in the final state and the reduction due to the @xmath50 initial state interaction ( see section [ mwhs ] ) . according to the above discussion the interaction in the three body final state @xmath261 is approximated by @xmath521 and @xmath522 , which denote the corrections for the coulomb interaction and the @xmath513 fsi , depending on the relative momenta @xmath523 and @xmath524 in the proton kaon and proton@xmath518 centre of mass systems , respectively . the coulomb correction is essentially given by the coulomb penetration factor @xmath223 of eq . @xmath525 with the relativistic coulomb parameter @xmath526 where @xmath62 is the fine structure constant , @xmath527 denotes the total energy in the @xmath528 subsystem and the triangle function @xmath101 is defined in eq . .the modification of the energy dependence close to threshold due to the @xmath513 fsi @xmath529 is determined by the relative momentum in the @xmath513 centre of mass system and the @xmath513 s wave phase shift @xmath530 , which has been expressed in terms of an effective range expansion in eq . , with @xmath531 and @xmath532 as spin averaged values of @xmath513 scattering length and effective range system can be in a spin singlet @xmath533 and spin triplet @xmath534 state , the amplitudes should be added incoherently . however , model calculations @xcite and experimental data @xcite suggest that the scattering parameters are rather similar for both spin states ( fig . [ pl_param ] ) . ] . following the result for the @xmath75 final state in eq . in section [ mwhs ] and using relation from section [ siv ] , the total cross section for the reaction @xmath535 writes as : @xmath536 where f is the flux factor of the colliding nucleons from defined in eq . using the energy dependence given in eq . an accurate description close to threshold is achieved , as shown in figure [ cross_etap_pkl]b in section [ mwhs ] ( see also @xcite ) . within meson exchange models , contributions to the reaction dynamics of the elementary strangeness dissociation are expected from direct production ( fig . [ pky_graph]a ) and from the exchange of nonstrange @xmath537 mesons including @xmath59 resonance excitation and @xmath538exchange in the strange sector ( fig . [ pky_graph]b ) . + pion and kaon exchange were suggested as mechanisms of the associated strangeness production in the nucleon nucleon interaction already by ferrari @xcite . as noted by laget @xcite , the contribution from direct production ( fig.[pky_graph]a ) to the total cross section is small but might be of relevance for differential observables as discussed below . modern parameterizations considering @xmath9 and @xmath10exchange reproduce the magnitude of the total cross section close to threshold within a factor of two or three ( grey dashed @xcite and grey dotted @xcite lines in fig . [ pkl_cross ] ) . in refs . @xcite the elementary @xmath539 and @xmath540 transitions are included phenomenologically induced strangeness production , either experimental values @xcite ( as in the work of laget @xcite ) or parametrizations ( cross sections determined from the resonance model approach @xcite used by li and ko @xcite , and sibirtsev @xcite ) , respectively , are employed . @xmath541 scattering data are available experimentally @xcite and in terms of phase shift analyses @xcite . ] and both exchange diagrams are found to be important in the case of the reaction @xmath535 , while @xmath542 production appears to be dominated by @xmath9exchange exchange mechanism in @xmath518 and @xmath543 production reflects the smallness of the @xmath544 coupling constant @xmath545 as compared to @xmath546 ( see e.g. @xcite ) ] . however , interference terms of the amplitudes as well as the effects of final state interactions have been neglected , although the latter have been shown to be of relevance in earlier studies @xcite . the same holds for a resonance model ( dashed line in fig . [ pkl_cross ] @xcite ) considering @xmath9 , @xmath11 and @xmath378exchange exciting baryon resonances up to @xmath547 coupling to the kaon hyperon channel ( fig . [ pky_graph]b left ) , i.e. the @xmath53 @xmath459 , @xmath24 @xmath548 , @xmath549 @xmath550 and @xmath27 @xmath551 resonances . the fully relativistic model essentially uses parameters fixed by the previous study on @xmath9 induced strangeness production @xcite and experimental branching ratios , interferences of the amplitudes are neglected . non resonant exchange diagrams , i.e. especially kaon exchange , are not considered from arguments based on the author s studies on @xmath552 reactions @xcite . as a result , @xmath9 exchange and excitation of the @xmath459 , coupling strongly to the @xmath351 channel , is found to give the dominant contribution close to threshold and is only exceeded by @xmath378 exchange at excess energies beyond @xmath553 . however , as fsi effects are neglected according to the authors the approach is neither expected nor intended to describe the close to threshold regime @xcite . within a factor of two , the total cross section data is described by an alternative approach @xcite considering @xmath9 exchange and excitation of the second @xmath53 resonance , the @xmath459 ( dotted lines in fig . [ pkl_cross ] ) in analogy to @xmath11 production arising from @xmath9 exchange and excitation of the first @xmath53 resonance ( @xmath367 ) . assuming the @xmath367 and the @xmath459 to dominate the @xmath211 and @xmath517 systems , respectively , leads to identical forms of the production operators and the spin angular momentum algebra . thus , after normalizing the scale of the @xmath554 excitation function to low energy data on the @xmath75 final state , the model gives an absolute prediction for @xmath518 production , while the energy dependence close to threshold is essentially determined by three body phase space modified by the @xmath513 final state interaction . the two corresponding curves shown in figure [ pkl_cross ] ( dotted lines ) differ in the choice of low energy @xmath513 scattering parameters @xcite determining the proton hyperon fsi ( see eq . ) . the situation is summarized by kaiser @xcite : separating the on shell s wave final state interaction and the invariant @xmath249 matrix at threshold , the information from the close to threshold total cross section can be `` condensed '' into essentially one number , the threshold transition amplitude parametrizing the @xmath249 matrix , with the energy dependence being accurately reproduced by three body phase space and fsi ( see also our discussion in section [ dopsmp ] ) . in an analysis based on tree level meson exchange diagrams pointlike @xmath390 exchange alone or the total vector @xmath448 , @xmath390 and @xmath555 exchange , are found to describe the transition amplitude used in this work by simple kinematical factors , i.e. @xmath556 , with @xmath40 denoting the mass of the second baryon @xmath557 in the final state and @xmath558 as threshold transition amplitude , i.e. @xmath559 and @xmath560 for @xmath9 and @xmath11 production in @xcite and @xmath561 in @xcite . ] . however , also pseudoscalar amplitudes , i.e. pointlike @xmath284 exchange alone , or in line with @xcite @xmath9 exchange followed by an excitation of the @xmath53 @xmath459 resonance when added to the total pseudoscalar and vector exchange contributions are in good agreement with the experiment . it is concluded @xcite that the unpolarized @xmath562 total cross section close to threshold does not provide enough information to determine the underlying production mechanism . instead , further information should be provided in the future by angular distributions and polarization observables , as reported at higher excess energies by the disto collaboration @xcite : the importance of @xmath10 exchange for @xmath518 production , which might already be expected from elementary amplitudes cross section is in the range of @xmath563 compared to @xmath564 for @xmath220 scattering ( see fig . [ kpkp_kk_sspspp]a ) . ] has been supported experimentally by data on the normal spin transfer coefficient @xmath565 in exclusive data at an excess energy of @xmath566 . @xmath565 is found to be large and the @xmath518 polarization correlated with the beam polarization is oriented opposite to the beam spin for forward @xmath518 production . based on the model described in @xcite , calculations by laget show @xmath565 to discriminate between the @xmath9 and @xmath10 exchange mechanism from a different spin coupling at the vertices between the polarized proton in the initial and the hyperon in the final state . while from angular momentum and parity conservation @xmath9 exchange leads to @xmath567 for forward production , kaon emission results in a spin flip ( @xmath568 ) and consequently , the latter is clearly favoured by the experimental result @xcite high energy ( @xmath569 ) data @xcite , which have been accounted for within quark model calculations @xcite . ] . recently , the negative sign of the spin transfer coefficient , i.e. the dominance of @xmath10 exchange for @xmath518 production , has been confirmed for two different beam momenta corresponding to excess energies of @xmath434 and @xmath570 , respectively @xcite . however , unexpectedly large negative values are also seen for backward production , where the hyperon is preferentially associated with the unpolarized target proton an observation , which is so far not reproduced within the meson exchange picture . the first close to threshold measurement of the @xmath518 recoil polarization at the cosy tof facility has shown a negative polarization for transverse momentum transfers @xmath571 @xcite polarization is also observed for @xmath572 in proton proton scattering at higher energies @xcite . ] , which implies evidence for s and p wave interference terms at the respective excess energy of @xmath573 according to @xcite . differential observables , like the occupation in a dalitz plot , or in projection invariant mass distributions of two particle subsystems , might also provide important information to identify the relevant production mechanism . a resonant production should influence for example the @xmath574 invariant mass distribution significantly at excess energies in the range of @xmath250 , as calculated within the resonance model approach of @xcite by sibirtsev et al . @xcite in @xcite . ] . first data has already been published @xcite and higher statistics distributions at several excess energies will be available from data taken at the cosy tof facility in the near future . @xmath518 production on a neutron target , i.e. via the reactions @xmath575 , might turn out to be sensitive on the relative strength of @xmath9 , @xmath378 and @xmath11 exchange , as it has been discussed in case of @xmath11 production @xcite . an important aspect in the theoretical interpretation of data close to the production threshold is the interplay between the type of hyperon nucleon interaction employed and the result of the model calculation . as already discussed by laget @xcite , with experimental observables being significantly influenced by the strong final state interaction at low transverse momenta , exclusive close to threshold data should be expected to eventually constrain strongly @xmath576 low energy scattering amplitudes . within a meson exchange approach , i.e. considering direct kaon emission as well as @xmath9 and @xmath10 exchange ( figs . [ pky_graph]a b ) and using an off shell @xmath249 matrix description for the hyperon nucleon interaction the sensitivity on the microscopic hyperon nucleon potential used has been studied in detail by kelkar and jain @xcite for @xmath518 production in comparison to the inclusive saturne data from @xcite . the authors conclude , that in order to separate different channels contributing at the @xmath577 threshold i ) exclusive data is needed ( which shall be discussed below ) to constrain hyperon nucleon interaction models and that ii ) an improved knowledge on the hyperon nucleon interaction is prerequisite to further detailed investigations on the relevant production mechanisms . we shall use these conclusions as guideline for the discussions during the next paragraphs . due to the short hyperon lifetimes direct @xmath578 scattering experiments are difficult to perform and , consequently , low energy scattering parameters are still rather unknown experimentally . in the case of the @xmath518proton system , experimental results have only been available at excess energies above @xmath579 @xcite . in contrast , a study in close to threshold production via the strong final state interaction covers an excess energy range down to threshold , otherwise inaccessible for elastic scattering experiments . assuming the energy dependence of the reaction @xmath535 close to threshold being determined by eq . , i.e. by three body phase space modified by the strong @xmath513 fsi and the coulomb repulsion in the @xmath516 subsystem , constraints on the spin averaged low energy @xmath580 scattering parameters have been extracted in @xcite from the differential occupation of dalitz plot distributions obtained from measurements at the cosy11 facility @xcite . however , in the analysis scattering length and effective range appeared strongly correlated , as inherent to the parametrization used in eqs . and . thus , only a combined fit using the complementary information of low energy @xmath580 elastic scattering experiments @xcite constrained by the dalitz plot analysis allowed to derive separate spin averaged values @xcite . the result new spin averaged values of @xmath580 s wave scattering length and effective range @xmath581 is shown in figure [ pl_param ] together with other experimental and theoretical values , where the extracted value of the @xmath580 scattering length is clearly fixed by the old data . although the fsi approach to low energy @xmath518proton scattering proves to be a powerful tool , it should be noted that the above values from @xcite have been derived assuming a cross section ratio according to the number of possible magnetic spin quantum numbers , i.e. a ratio of three to one between the triplet and singlet @xmath580 final states , which was confirmed in the early bubble chamber experiments . consequently , the analysis implies that the underlying reaction mechanism in proton proton scattering does not favour either the singlet or triplet state , different to the assumed ratio of spin states . future threshold production experiments employing both polarized beam and target should dispel this concern and allow to extract singlet and triplet hyperon nucleon scattering parameters separately . eventually referring to the first conclusion of kelkar and jain @xcite cited above , further constraints on the present understanding of close to threshold hyperon production have arisen from data on the reaction @xmath582 taken at the cosy11 installation @xcite : at equal excess energies up to @xmath583 with respect to the hyperon production thresholds the total cross sections have been determined for both @xmath518 @xcite and @xmath543 @xcite production . a fit to the data including the hyperon nucleon fsi is shown in figure [ pkls_cross ] and has been performed using the parametrization of the energy dependence developed by fldt and wilkin @xcite as given by eq . , yet neglecting the weak variation of the cross section due to the variation of the flux factor f. it should be noted that gasparyan @xcite concludes that the proton hyperon fsi is less important for the @xmath577 rather than for the @xmath584 channel . a similar result gives the preliminary analysis of new data taken at cosy11 for both reaction channels up to excess energies of q = 60 mev @xcite . the most striking feature of the data is the observed @xmath543 suppression in the energy range close to threshold with @xmath585 while at excess energies @xmath586 this ratio is known to be about 2.5 @xcite . considering only @xmath9 exchange , data on @xmath9 induced hyperon production via @xmath587 result in a ratio of @xmath588 @xcite , clearly underestimating the experimental value of relation . kaon exchange essentially relates the ratio @xmath589 to the ratio of coupling constants squared @xmath590 . although there is some uncertainty in the literature on their values @xcite , a @xmath591 production ratio of 27 follows from the suitable choice of the su(6 ) prediction @xcite , in good agreement with experiment . however , effects of final state interaction as well as the importance of @xmath9 exchange especially for @xmath543 production are completely neglected by this simple estimate . inclusive @xmath284 production data taken at the spes 4 facility at saturne in proton proton scattering at an excess energy of @xmath592 with respect to the @xmath593 threshold show enhancements in the invariant mass distribution at the @xmath580 and @xmath577 thresholds of similar magnitude @xcite . however , with only the kaon being detected , it is not clear whether the enhancement at the @xmath577 threshold is due to @xmath542 production . qualitatively , a strong @xmath594 final state conversion might account for both the inclusive results from spes 4 as well as the @xmath543 depletion in the cosy11 data . in production experiments , evidence for @xmath594 conversion effects is known experimentally from exclusive hyperon data via @xmath595 @xcite ( for a theoretical interpretation see @xcite ) . in hypernuclear physics , the strong @xmath596 mode appears as the dominating decay channel of @xmath542hypernuclei @xcite , and the @xmath577 interaction reflects in both shift and broadening of the last x rays observed from @xmath597 atoms @xcite . theoretically , modern hyperon nucleon interaction models show a significant cusp effect for the @xmath584 cross section at the @xmath577 threshold @xcite . in exploratory calculations performed within the framework of the jlich meson exchange model @xcite , taking into account both @xmath9 and @xmath10 exchange diagrams and rigorously including fsi effects in a coupled channel approach , a final state conversion is rather excluded as dominant origin of the observed @xmath543 suppression . while @xmath518 production is found to be dominated by kaon exchange in agreement with the disto results at higher excess energies discussed above both @xmath9 and @xmath10 exchange turn out to contribute to the @xmath543 channel with similar strength . it is concluded by gasparian et al . @xcite , that a destructive interference of @xmath9 and @xmath10 exchange diagrams might explain the @xmath543 suppression close to threshold and a good agreement with the cosy11 data is obtained after including the overall reduction @xmath262 in eq . due to the @xmath50 interaction in the initial state @xcite . an experimental study of @xmath542 production in different isospin configurations should provide a crucial test of the above interpretation : for the reaction @xmath598 an opposite interference pattern is found , i.e. the @xmath599 channel is enhanced for a destructive pattern as compared to the constructive interference of @xmath10 and @xmath9 exchange . for the choice of a destructive interference favoured by the @xmath600 channel , the corresponding ratio @xmath601 is in good agreement with high energy data , whereas the alternative choice is rather difficult to reconcile with existing data at higher energies @xcite . in contrast , for the reaction @xmath602 the same interference pattern occurs as for the @xmath600 channel , resulting in a ratio @xmath603 close to threshold for either choice @xcite , while a ratio close to unity is found in the literature for higher energies @xcite . as stated by the authors of ref . @xcite , contributions from direct production ( fig . [ pky_graph]a ) as well as heavy meson exchanges ( fig . [ pky_graph]b ) have been neglected so far in these calculations but might influence the @xmath591 production ratio ( see also @xcite ) . in fact , employing both a @xmath9 and @xmath10 exchange based meson exchange model without any interference of the amplitudes following @xcite and the resonance model of @xcite the close to threshold @xmath591 production ratio has been studied in @xcite . within a factor of two of the experimental error bars in @xcite both models with parameters fixed at data from excess energies @xmath604 are in reasonable agreement with the data , i.e. according to the authors the total cross section data close to threshold are not sensitive on the details of the model used @xcite and @xmath10 exchange amplitudes are concluded by sibirtsev et al . @xcite to contribute to both @xmath518 and @xmath543 production with similar magnitude . as mentioned in @xcite , an experimental determination of the kaon exchange contribution close to threshold e.g. from polarization observables as in @xcite should be crucial to identify the dominant reaction mechanisms . ] . the latter study has been critically discussed in @xcite , where the strangeness production is modeled in an effective lagrangian approach following @xcite and strangeness production proceeds via @xmath9 , @xmath378 , @xmath390 and @xmath26 exchange exciting the nucleon s wave resonance @xmath459 and the p wave resonances @xmath548 and @xmath550 . for both the @xmath518 and @xmath543 production channel , @xmath9 exchange followed by an excitation of the @xmath459 is found to dominate the total cross section close to threshold ( see also @xcite ) , though an on shell coupling of the @xmath459 to the @xmath605 channel is suppressed with the channel opening at @xmath606 . in contrast , the influence of the @xmath459 on @xmath543 production and possible interference effects of resonance contributions have been neglected in @xcite . at larger excess energies ( @xmath607 ) the @xmath548 is concluded in the work of shyam et al . @xcite to dominate both reaction channels , which would be expected when considering coupling constants only production in the resonance model of @xcite proceeds preferrably via @xmath9 and @xmath11 exchange , the latter being enhanced for the @xmath543 compared to the @xmath518 channel due to the large branching ratio of the @xmath548 to the @xmath368 channel . however , as pointed out in @xcite the @xmath11 exchange contribution is tainted with rather large uncertainties due to different values of the @xmath338 coupling constant extracted from experimental data ( see @xcite and section [ mwhs ] ) . ] . it is worth noting , that the obe calculations performed by laget @xcite , although not using the interference of pion and kaon exchange graphs explicitely and just choosing the relative sign to maximize the cross section , not only describe the recent close to threshold data on the @xmath591 ratio within a factor of two , as well as the polarization transfer results of the disto experiment @xcite , but also give an accurate description of the @xmath576 invariant mass distributions obtained in the inclusive measurements at saturne @xcite cited above : in particular , the fsi contribution from the direct @xmath608 amplitude is found to significantly enhance the kaon spectrum just above the @xmath580 threshold , while the strong @xmath609 coupling `` induces only a small dip close to the @xmath577 threshold '' in the @xmath9 and @xmath10 exchange amplitudes @xcite . on the other hand , it is this strong channel coupling and the resulting rapid variation of the @xmath534 @xmath584 amplitude near the @xmath577 threshold , which appears responsible for the sharp rise of the invariant mass distribution obtained at spes 4 at the @xmath577 threshold via the `` kaon direct emission amplitude '' , i.e. the coupling of one proton to a @xmath610 pair and a following interaction with the second proton ( fig . [ pky_graph]a ) . recently , measurements of the @xmath591 production ratio in proton proton collisions have been extended up to excess energies of @xmath611 at the cosy11 installation @xcite , in order to study the transition between the low energy enhancement of the ratio reported in @xcite and data at excess energies larger than @xmath612 . in comparison to the experimental data , in figure [ pkls_ratio ] calculations are included obtained within the approach of gasparian et al . @xcite assuming a destructive interference of @xmath9 and @xmath10 exchange with different choices of the microscopic hyperon nucleon model @xcite . as it was already emphasized in @xcite , the result depends on the details especially the off shell properties of the microscopic hyperon nucleon interaction employed , although the actual choice does not alter the general result in @xcite of only the destructive interference of @xmath10 and @xmath9 exchange explaining the experimentally observed suppression of the @xmath543 signal close to threshold . at the present stage of theoretical investigations , both the good agreement found for jlich model a @xcite with the close to threshold result for @xmath589 in relation [ rlamsig_eq ] and for the nijmegen model ( solid line in fig . [ pkls_ratio ] @xcite ) with the energy dependence of the cross section ratio in figure [ pkls_ratio ] should rather be regarded as accidental . in the latter case an su(2 ) breaking in the @xmath534 @xmath577 channel had to be introduced @xcite . consequently , the relation between the @xmath613 amplitude with the @xmath614 and @xmath615 channels becomes ambiguous . only one of the choices leads to the good agreement with the data shown in figure [ pkls_ratio ] ( solid line ) , whereas the other one results in a completely different prediction @xcite . calculations using the new jlich model ( dotted line in figure [ pkls_ratio ] @xcite ) do not reproduce the tendency of the experimental data . it is suggested in @xcite that neglecting the energy dependence of the elementary amplitudes and assuming s waves in the final state might no longer be justified beyond excess energies of @xmath293 . conclusions from figure [ pkls_ratio ] , as well as considerations by kelkar and jain @xcite for the @xmath535 reaction clearly demonstrate , that once the reaction mechanism for close to threshold hyperon production is understood exclusive hyperon production data should provide a strong constraint on the details of hyperon nucleon interaction models . finally , in future high resolution hyperon production experiments close to threshold may give further information on the possible existence of low lying strange dibaryons . narrow , stable six quark objects @xmath616 and @xmath617 with a @xmath618 cluster substructure and strangeness @xmath619 coupling to p wave nucleon hyperon systems have been postulated from an extended mit bag model @xcite below and above the @xmath577 thresholds with widths in the range of @xmath620 to @xmath73 . experimental signatures consistent with a strange dibaryon hypothesis have been found both in strangeness transfer reactions @xmath621 @xcite and in the inclusive @xmath622 data taken at saturne @xcite at excess energies of 175 and @xmath623 with respect to the @xmath613 threshold . however , the statistical accuracy in these experiments turned out to be too small for unequivocal conclusions . for a high resolution study of the inclusive @xmath624 reaction channel new experimental efforts are on their way at the big karl spectrometer at cosy , with a resolution improved by a factor of five and a higher statistical accuracy compared to the saturne measurements @xcite . single and double pion production in elementary @xmath351 , @xmath625 and @xmath12 reactions is an important source to gain information on nucleon nucleon , nucleon pion and pion pion interactions . apart from these fundamental aspects , the pion production on nucleons also enables the study of the properties of nucleonic resonances , which might be excited in the reaction processes . the reaction @xmath626 for instance has been found to be influenced by the formation of the @xmath627 resonance @xcite with a subsequent decay into @xmath628 . therefore , studies on this reaction channel allow to gain information about the @xmath629 decay amplitudes . additionally , contributions of the @xmath25 resonance , followed by its decay @xmath630 , play a non negligible role in the near threshold production of pion pairs via the reaction channels @xmath631 and @xmath632 @xcite . in case of the pion pair production in the nucleon nucleon scattering similar contributions of nucleonic resonances are under discussion . recently , detailed calculations on the two pion production in nucleon nucleon interactions have been performed @xcite , considering non resonant terms as well as contributions from nucleonic resonances with subsequent decays into a nucleon and pions . one of the main results of these studies is the prediction of a dominant effect of the @xmath25 roper resonance , followed by its decay @xmath633 , in the region of low excess energies . according to this , the absolute scale of the low energy total cross section of reaction channels , where the pions can be in an isospin zero state , e.g. @xmath634 and @xmath635 , should be determined by the excitation of this resonance . in this region of excess energies contributions of non resonant terms are discussed to be negligible for the description of the excitation function of both these reaction channels . however , at higher energies ( @xmath636 ) the excitation of @xmath20 resonances has been evaluated to dominate the total cross sections . a different situation is given in the case of the @xmath637 reaction channel . here contributions of the roper resonance with the subsequent decay into @xmath638 are forbidden by conservation laws and contributions of remaining @xmath59 production diagrams are suppressed at low excess energies . since the non resonant terms are also negligible in this case , the dominating processes are given by @xmath20 resonance excitations . it is worth noting that the discussed model calculations yield predictions for the excitation functions for different two pion production reactions from threshold up to high excess energies ( @xmath639 ) and all are based on the same feynman diagrams . however , especially in the close to threshold region the data situation for most double pion reaction channels is far from being optimal . recently , three isospin independent reaction channels on the two pion production in proton proton interactions , @xmath640 , @xmath635 and @xmath641 , have been studied in high statistics precision measurements at the promice / wasa facility at celsius and resulted in the first total and differential cross sections in the near threshold region @xcite . in fig . [ twopiall]a the total cross sections for the @xmath481 production in proton proton scattering are presented up to an excess energy of @xmath639 . the solid and the dashed lines represent the above discussed predictions of ref . @xcite for this reaction channel and correspond to two possible solutions in the applied model . obviously these calculations are in good agreement with the experimental data both close to threshold and in the region of higher excess energies , while the model seems to underestimate the data at intermediate energies . it should be noted that the data points from dakhno et al . @xcite have been obtained using a deuterium filled bubble chamber and selecting events from the quasi free two pion production @xmath642 . however , in the analysis of the data the fermi motion of the nucleons in the deuteron has been neglected . in later calculations an adequate consideration of the deuteron wave function has been found to shift the effective excess energy by @xmath643 towards higher energies @xcite , corresponding to a shift of the data towards the region expected by the discussed model calculation . figure [ twopiall]b displays the experimental situation for the reaction channel @xmath644 . the new data from promice / wasa disfavour the solution of the model calculation indicated by the solid line but are in very good agreement with the prediction shown by the dashed line . however , the high energy data from ref . @xcite are in disagreement with both curves and are clearly overestimated . finally , the close to threshold total cross section data of the reaction channel @xmath641 are presented in fig . [ twopiall]c.different to the previous situation now the data is underestimated by both solutions of the model calculations , especially in the near threshold region . since now a production via the @xmath645 is forbidden , this observation might indicate contributions from processes like @xmath646 larger than predicted by theory . it should be mentioned that contributions of higher partial waves and final state interactions have been neglected in the model calculations , indicated by the presented curves . however , the effect of the @xmath50 fsi has been investigated and is reported to increase the close to threshold cross sections by nearly one order of magnitude , while at @xmath647 this effect is reduced to an increase of a factor of @xmath648 @xcite . according to this , considering the interaction of the final state baryons is expected to reduce the apparent discrepancy between the @xmath649 total cross section data and the corresponding model calculations . surprisingly , in case of the @xmath481 and @xmath480 pair production , the near threshold data are described reasonable by these calculations neglecting the @xmath50 fsi . furthermore , signals for the contribution of higher partial waves are reported for the near threshold data @xcite , which also have been neglected in the discussed model calculations . this observation might be interpreted as a signal of a heavy meson exchange ( @xmath26 , @xmath378 ) between the interacting protons . it is well known and cited in the literature as the abc anomaly @xcite that the missing mass enhancement at around @xmath650 with a width of @xmath475 observed in the inclusive measurements of the @xmath651 reaction at a proton beam energy of @xmath652 ( corresponding to an excess energy of @xmath653 for the @xmath654 production ) is not excited in the @xmath655 final state . detailed investigations confirm that this observation is to be associated with the isospin wave @xmath654 double pion state but additionally stress that the missing mass peak and its width vary with the beam momentum @xcite . further , since the isoscalar @xmath654 scattering length is small @xcite the observed anomaly is supposed to be of kinematical origin and associated with the intermediate excitation of two @xmath20 resonances @xcite . recently the momo ( * m*onitor*o*f*m*esonic*o*bservables ) collaboration at cosy investigated the @xmath471 production closer to threshold in an exclusive measurement of the @xmath656 reaction at a beam energy of @xmath657 ( excess energy @xmath658 for the @xmath654 production ) , where the abc enhancement is expected to show up near the centre of the available @xmath471invariant mass distribution @xcite . in figure [ jahn_momo_fig03]a d differential cross sections in terms of four kinematical variables are presented : a ) the @xmath481 excitation energy @xmath659 , where @xmath660 is the two pion invariant mass.b ) the @xmath9@xmath661 excitation energy.c ) the angle ( @xmath662 ) between the proton and one of the @xmath9 s in the overall centre of mass ( c.m . ) system.d ) the angle ( @xmath663 ) between the relative @xmath664 momentum and the beam axis in the overall c.m . system . the @xmath665 distribution at this near threshold data is pushed closer to the maximum values of excitation energy and is in contrast to the original abc result @xcite which showed an enhancement over phase space at @xmath666 . the distribution of the @xmath9@xmath661 excitation energy is rather consistent with the phase space prediction . the significant anisotropy of @xmath662 confirms the importance of higher partial waves involved in the reaction process . the momo collaboration concludes @xcite that the agreement between the @xmath665 data of fig . [ jahn_momo_fig03]a with the solid line ( phase space multiplied by @xmath665 ) and the fair linearity of @xmath667 indicates that the @xmath481 system is essentially produced with an internal angular momentum @xmath668 . on the other hand , as outlined in section [ dppinns ] , the valencia group @xcite developed a model for the two pion production resulting in a dynamical origin for the small invariant masses of the low energy part , where the two pions can be in an isospin zero state . here the reaction dynamics is dominated by the intermediate excitation of the @xmath59(1440 ) roper resonance followed by the @xmath669 decay . the experimental situation seems to be the following : s wave @xmath471 production has been observed in the @xmath656 reaction very close to threshold @xcite and at high excess energies ( @xmath670 ) @xcite whereas at an intermediate excess energy range a p wave @xmath481 production process dominates . similar results have been obtained in @xmath671 @xcite and in the @xmath672 @xcite reactions . after the remarkable results of p wave production in the @xmath481 system the momo collaboration continued equivalent measurements for the @xmath673 reaction at excess energies of @xmath674 , @xmath290 and @xmath675 . the @xmath487 invariant mass spectra in units of the @xmath487 relative energy @xmath676 are shown in figure [ jahn_kk_momo ] . a clear signal of the @xmath45meson production is seen above the continuous non resonant @xmath487 spectrum . the dashed lines in all three parts of the figure correspond to simple phase space calculations , where the q dependences of the cross sections ( @xmath677 and @xmath678 ) were taken to scale as @xmath679 for the three body and @xmath680 for the two body final state , respectively . the relative angular momentum between the two kaons was assumed to be @xmath681 . no significant deviation of the data from these summed phase space distributions is observed and thus it is interesting to note , that no indication of p wave in the @xmath487 system is present in the mass spectra and angular distributions , in contrast to the above discussed data for the two pion production via the reaction @xmath656 . measurements on the @xmath426 meson pair production are of considerable interest in the context of the continuing discussion on the nature of the scalar resonances @xmath463 and @xmath464 , as already discussed in section [ ssec ] . available close to threshold data @xcite on the elementary antikaon production channel in proton proton scattering , @xmath683 , are shown in figure [ kk ] . in comparison , parametrizations of the energy dependence of the total cross section based on a four body phase space behaviour and for a resonant production via the @xmath473 are included . obviously , within the experimental error bars and neglecting the possible influence of higher partial waves production at an excess energy of @xmath684 is still consistent with isotropic emission . ] , the data is consistent with both a non resonant and resonant production , as already concluded from figure [ mm_kk_f0 ] in section [ ssec ] , i.e. from the @xmath487 invariant mass distribution obtained at an excess energy of @xmath685 @xcite . furthermore , exclusive @xmath83 production data are of special interest with respect to subthreshold kaon production experiments in nucleus nucleus interactions , which are expected to probe the antikaon properties at high baryon density . recent inclusive subthreshold measurements @xcite resulted in comparable @xmath284 and @xmath83 yields at the same energy per nucleon below the production thresholds for the elementary reactions @xmath686 , which has been discussed as an indication for a possible restoration of chiral symmetry ( as reviewed in @xcite ) . available inclusive elementary @xmath284 and @xmath83 total cross sections in nucleon nucleon interactions are shown in figure [ kincl ] . in contrast to the results obtained in subthreshold measurements , in proton proton scattering the inclusive @xmath284 cross section exceeds @xmath83 production by more than an order of magnitude in the near threshold region . to explain this observation , different models @xcite consider kaons ( @xmath687 ) and antikaons ( @xmath688 ) to be subject of repulsive and attractive forces within the nuclear medium , respectively . recent theoretical studies using a chiral unitary approach to describe the @xmath689 s wave interaction allow for a self consistent microscopic implementation of medium effects @xcite . the @xmath83 self energies obtained are consistent with kaonic atom data , as shown in @xcite . near threshold meson production experiments in nucleon nucleon interactions might help to obtain fundamental information on the elementary production processes and final state interactions . extending those measurements from the binary nucleon scattering to reactions with three or four participating nucleons , i.e. meson production in the proton deuteron or deuteron deuteron scattering , allows to study the role of further participating nucleons . this facilitates a gain in information on possible multi step production processes , which are naturally absent in the nucleon nucleon case . of considerable interest is the production of @xmath11 mesons in the @xmath690 reaction . studied at the saturne accelerator in the near threshold region @xcite this channel is reported to expose remarkable features . in spite of the much higher momentum transfer , the observed threshold amplitude @xmath691 for the @xmath11 production , defined by the unpolarized centre of mass cross section @xmath692 was found to be comparable to that for the @xmath693 reaction at its respective threshold @xcite . in addition , while being consistent with s wave production , @xmath694 decreases by a factor of three from threshold up to a centre of mass momentum of emitted @xmath11 mesons of @xmath695 ( @xmath696 ) . this observed strong decrease in the near threshold region can be addressed to a strong @xmath697 final state interaction and is therefore of special interest in view of the existence of @xmath11mesic nuclei . although being the topic of several theoretical investigations @xcite , the possibility for the formation of quasi bound @xmath698 states is still an open question . @xmath699 + the experimental situation for the @xmath700 reaction is presented in fig . [ he3rev ] , displaying total cross section data from threshold up to an excess energy of @xmath701 . the saturne data @xcite ( circles ) expose an energy dependence , which clearly differs from pure phase space expectations : according to naive two body phase space considerations , neglecting higher partial waves , an excitation function according to @xmath702 is expected . one promising ansatz to describe the reaction process was given by kilian and nann @xcite , taking into account both nucleons from the target and considering double scattering diagrams . a sketch of such a double scattering diagram is illustrated in the left part of fig . [ he3etafey ] . calculations by fldt and wilkin @xcite are in line with this approach and describe the @xmath703 reaction as an initial @xmath704 ( @xmath124 ) production followed by a final reaction according to @xmath705 ( @xmath706 ) . these investigations lead to a prediction of the spin averaged threshold amplitude squared of @xmath707 , underestimating the measured results by a factor of 23 . since none of the considered sub reactions exposes a significant variation of the cross section in the relevant energy region , the observed energy dependence of the production amplitude @xmath691 has been interpreted to be a signal of a strong final state interaction of the @xmath11nucleus system . introducing the @xmath698 final state interaction by an s wave scattering length formula @xmath708 with @xmath709 as the threshold amplitude , the best description of the observed energy dependence of the saturne total cross section data is achieved using a scattering length of @xmath710 , as can be seen in fig . [ he3rev ] ( solid line ) with the absolute scale of the curve being scaled to fit the data . such a large scattering length might be associated with the existence of a quasi bound @xmath698 state . + in the right part of fig . [ he3rev ] the extracted production amplitude squared @xmath711 is presented as function of the centre of mass momentum of the ejectiles . it should be noted that in the close vicinity of the production threshold and with a given cross section , the production amplitude squared @xmath711 depends crucially on the momentum @xmath712 and , therefore , on the exact masses of the ejectiles . here a mass of the @xmath713nucleus of @xmath714 @xcite and a meson mass of @xmath715 @xcite have been applied . while a constant amplitude would imply a behaviour according to pure s wave phase space considerations , the observed shape exposes a strong momentum dependence and can be reproduced well by the discussed formalism up to excess energies of @xmath716 , corresponding to @xmath717.however , recent measurements at higher excess energies from the gem collaboration at cosy @xcite and from the wasa / promice collaboration at celsius @xcite clearly deviate from the predicted energy dependence of the cross section ( fig . [ he3rev ] ) . moreover , different to the saturne results , the centre of mass angular distribution of the emitted @xmath11 mesons strongly differs from an isotropic emission and exposes a peaking in forward direction . motivated by results on the @xmath11 meson production in @xmath718 and @xmath50 interactions , another approach to describe the production process has been evaluated @xcite where the @xmath11 meson creation is described by the excitation of the @xmath53 @xmath367 resonance as shown in figure [ he3etafey ] ( right ) . neglecting effects of other resonances , the production amplitude can be expressed by a breit wigner distribution @xcite : @xmath719 with @xmath720 as the mass of the @xmath53 resonance and @xmath721 as the corresponding width . since the production threshold is located in the vicinity of the resonance mass @xmath720 , the energy dependence of the width @xmath722 has to be considered : @xmath723 @xmath699 + the parameters @xmath724 indicate the branching ratios for different decay modes of the resonance while the centre of mass momenta of the @xmath11 and @xmath9 mesons are represented by @xmath725 . a corresponding calculation , scaled to fit the gem data point , is given in figure [ he3rev ] by the dash dotted line . obviously , the prediction of this approach fails to describe the shape of the saturne data . this observation might indicate an influence of a strong @xmath698 final state interaction only in the close vicinity of the production threshold . alternatively , this effect might be caused by a transition between different production processes . + concluding , the @xmath703 reaction appears to be an interesting and important channel to study both the meson nucleus final state interaction as well as reaction processes involving three nucleons . however , the body of data is far from being optimal , especially in the near threshold region . new data from promice / wasa and cosy11 is currently under evaluation and will improve the experimental situation . from the theoretical point of view the existence of mesic nuclei or quasi bound meson nucleus systems is not excluded , however , up to now a compelling experimental proof for the formation of such a state is still missing . it is obvious that the possibility to create such states crucially depends on both the properties of the meson and the nucleus as well as on the sign and the strength of the meson nucleus interaction . since the @xmath9nucleon and @xmath284nucleon interaction have been found to be repulsive , it is unlikely that they will form quasi bound states . contrary , there are evidences for an attractive @xmath83-nucleon interaction , however , corresponding experiments would suffer from the low production cross sections for @xmath83 mesons . furthermore , the coulomb interaction is expected to screen corresponding physical observables and might lead to the formation of mesonic atoms bound by the coulomb interactions . different to this , the @xmath11 meson is uncharged and the observation of an attractive @xmath11nucleon interaction led to speculations concerning the existence of @xmath11nuclear quasi bound states . in the absence of @xmath11meson beams such states bound by the strong interaction would offer a new possibility to study the @xmath11nucleon interaction since the meson would be trapped for a relatively large time in nuclear medium . predicted by haider and liu @xcite , detailed studies on the formation of such states have been the topic of recent investigations @xcite , resulting in different lower limits of the atomic number a for which bound states should exist . while e.g. in @xcite the lower mass number is estimated in the framework of an optical model approach to be @xmath726 , there are considerations that even for nuclei of lower masses ( @xmath727 ) bound systems might exist @xcite , see also section [ mwhs ] . experimental evidences for bound states have been found in the near threshold production of @xmath11 mesons in the reaction channel @xmath728 . the unexpected large production amplitude as well as its rapid decrease with increasing energy ( figure [ he3rev ] ) is attributed to a strong s wave final state interaction associated with a large @xmath11@xmath661 scattering length @xcite.such a large scattering length ( see fit in figure [ he3rev ] ) might imply the existence of a quasi bound state . assuming the @xmath698 to form a quasi bound state , it is therefore interesting to compare @xmath703 production data with the ones of the reaction @xmath729 , where according to the higher mass number , the formation of a bound state should be even more probable . a comparison of corresponding near threshold data is presented in figure [ he34 ] . a fit to the data using the scattering length formula provides information on the scale of both the real and imaginary part of the scattering length of the meson nucleus system but is unsensitive to the sign of the real part of @xmath730 . however , information on this sign can be obtained using a lowest order optical potential @xcite : @xmath731 with the @xmath11nucleon reduced mass @xmath732 , the mass number of the nucleus a and the @xmath11nucleon scattering length @xmath733 . using gaussian nuclear densities with rms radii of @xmath734 ( @xmath735 ) for the @xmath736 ( @xmath737 ) system , meson nucleus scattering lengths of @xmath738 with negative real parts for both systems have been obtained . fits to existing close to threshold data for both reaction channels using these scattering lengths and eq . are presented in figure [ he34 ] . obviously , the data are reproduced well by this formalism . these results imply the existence of quasi bound meson nucleus states in the systems @xmath736 and @xmath737 , however , unambiguous evidence is still missing . therefore , detailed investigations on both reaction channels as well as on reactions using heavier targets are highly recommended . conservation laws restrict the quantum numbers for the production of mesons in the nucleon nucleon interaction . here especially the pion production is of considerable interest since it exhausts all inelasticity up to about @xmath739 beam momentum in the nucleon nucleon collision.as outlined in references @xcite the conservation of isospin allows the total cross section for pion production to be expressed as an incoherent sum of three independent partial cross sections given by : @xmath740 , @xmath741 and @xmath742 , where the first ( second ) index indicates the isospin of the initial state ( final state ) two nucleon system and @xmath743 is ruled out due to isospin conservation arguments.the three cross sections can be determined via the reactions : [ cols="<,^,<,^,<,^,<,^ " , ] the choice of a certain target material for accelerator experiments as well as its chemical and mechanical properties is strongly correlated with the given experimental conditions and , certainly , with the reactions which should be investigated . since in this article we concentrate on unpolarized meson production experiments off nucleons and deuterons , here we will restrict ourselves on corresponding targets . + proton and deuteron targets are commonly provided using hydrogen and deuterium as raw material . in the absence of free neutron targets deuterons can be used as substitute , considering the weakly bound deuterons as effective neutron targets and restricting on reactions with the bound protons acting only as spectator particles . for the sake of completeness we want to point out that in this case the fermi motion of the nucleons has to be considered , especially in the regime of low excess energies . + near threshold production experiments are commonly connected with low cross sections in the order of picobarns or nanobarns . to obtain sufficient counting rates during experimental runs , high luminosities in the order of @xmath744@xmath745s@xmath746 are necessary and can be achieved using targets of high areal density and intense accelerator beams . neglecting technical limitations like space charge effects , from the experimental point of view a combination of high intense ion beams and thin targets are preferable since in this case the quality of the cooled accelerator beam is nearly uninfluenced during the passage of the target , granting well defined kinematical conditions as well as single particle interactions . in practice one has to distinguish between internal and external beam experiments . in the latter case a well prepared and extracted accelerator beam is shot onto a target and beam particles not interacting with the target are lost . therefore , liquid or solid state targets of sufficient high areal density have commonly to be used in order to obtain the required high luminosities . contrary , in internal beam experiments carried out at storage rings the circulating beam particles pass the target material 10@xmath747 to 10@xmath748 times per second and projectiles not interacting with the target remain in the beam . therefore , to achieve similar luminosities compared to external beam experiments only thin targets are necessary moreover , to preserve the quality of the circulating beam and to obtain life times of the accelerated beam up to hours , effective areal target densities below @xmath749 atoms/@xmath750 are desirable and can be achieved using gas and cluster jets , pellets and thin fibers . considering hydrogen or deuterium as raw material such targets can be realized in different ways . fiber targets with diameters of only a few micrometers are of special interest for storage ring experiments with the request of high target densities in combination with a point - like target and uhv conditions . due to the relatively simple target installation , 4@xmath9 detection systems are possible . using carbon for example as target material , fiber diameters down to 5@xmath193 m are accessible , corresponding to @xmath751 atoms/@xmath752 @xcite and allowing to compensate beam heating effects by beam cooling devices . fiber targets for interactions off protons and deuterons naturally consist of chemical compounds like propylene such as the target for the internal beam experiment edda at cosy , where propylene fibers with diameters of @xmath753 m are used . however , apart from those advantages one has to keep in mind that the mass of such ch@xmath754 fibers is mostly given by `` carrier '' material carbon , which influences the performance of the data aquistion system and , moreover , causes a physical background in the data which has to be considered . liquid hydrogen and deuterium targets with their high volume densities ( @xmath755 0.07 g/@xmath756 ) are reasonable for the use at external beam experiments . depending on the experimental requirements large target volumes even up to several liters are accessible . however , precision experiments on the near threshold meson production force to use only small target volumes in the interaction region in order to minimize energy losses and effects of multiple scattering in the target material . an example of an experimental facility using a thin liquid hydrogen target is the cosy tof installation @xcite . here hydrogen is liquefyed using a commercial cryogenic cold head , providing a compact target volume of several millimeters in diameter . since such targets naturally have to be encased in special vessels , the choice of appropriate entrance windows is of great importance to minimize the physical background . at cosy tof for example thin windows of only @xmath757 1 @xmath193 m mylar foil with a diameter of @xmath757 4 mm are routinely in use , granting good signal to background conditions . pellet target installations can be used to prepare streams of frozen gas micro - spheres as targets for internal beam experiments . similar to fiber targets , pellet targets can be associated to the class of the solid state targets and allow to provide high areal densities in combination with point - like interaction regions , well suited for 4@xmath9 detector geometries . moreover , contrary to fiber targets these installations allow to produce pure hydrogen or deuterium pellets without any carrier material , however , at the cost of an increased experimental effort and vacuum conditions of @xmath758 mbar in the scattering chamber . at the celsius ring in uppsala such a pellet target was build up for the wasa experiment @xcite . the pellets are produced by pressing liquid hydrogen at triple point conditions ( 14 k , 72 mbar ) through a glass nozzle of 20 @xmath193 m diameter . by means of a piezo - electric transductor connected to the nozzle the stream of hydrogen is broken into uniformly sized micro - spheres ( pellets ) as shown in figure [ pellets ] . this device was designed to provide a stream of hydrogen pellets with diameters of 50 @xmath193 m at a repetition rate of @xmath757 68 khz . while the pellets itself are locally very thick ( @xmath759 atoms/@xmath752 ) , the effective areal density in the scattering chamber is in the order of 10@xmath760 atoms/@xmath752 due to the minimized geometrical beam - target overlap . supersonic gas jet beams can be used as targets for both internal and external beam experiments . produced by feeding laval - nozzles with gases like hydrogen or deuterium , high areal target densities ( @xmath761 atoms/@xmath752 ) can easily be achieved close to the nozzle exit @xcite . similar to pellet targets , proton or deuteron targets without effects of windows or carrier materials can be provided by operating the installation with gases of high purity . however , since the volume density of a gas jet beam drastically decreases both in longitudinal and transversal direction , the nozzle has to be placed very close to the interaction point in the scattering chamber . therefore , the advantage of high target densities of such installations is commonly connected with higher gas loads in the scattering chamber and a limited spatial resolution of the interaction point if no vertex detector is used . the great advantage of such targets is the possibility to use almost all gases as target material and to adjust the areal density by orders of magnitude by simply changing the gas input pressure . similar to conventional gas jet targets , cluster beams @xcite can be produced by using fine nozzles with a convergent - divergent shape ( laval - nozzle ) and , therefore , almost all gases can be used . operated at appropriate low temperatures , a partially condensation of the used gas is possible and clusters with sizes of 10@xmath762 to 10@xmath748 can grow . separated from the surrounding conventional gas beam by a set of collimators ( see fig . [ cluster ] ) , cluster beams with a homogeneous volume density , a sharp boundary , a small angular divergence and well defined beam dimensions can be prepared for interaction with circulating accelerator beams @xcite . cluster beam installations are frequently used at cooler rings . using laval - nozzles with diameters of 11 - 16 @xmath193 m , low nozzle temperatures of @xmath75725 k and gas input pressures of @xmath75718 bar , hydrogen and deuterium cluster beams with areal densities in the order of @xmath763 atoms/@xmath752 in combination with uhv conditions are accessible in the interaction region . furthermore , similar to gas beam targets the areal density can easily be adjusted over orders of magnitude by changing the gas input pressure or the nozzle temperature @xcite . special efforts are made for an optimal detector design to match the required conditions for kinematically completeness and undisturbed determination of the four momentum vector of the ejectiles . a kinematically complete event is characterized by knowing all four momentum vectors of the participating particles . the experimental information is limited by : i ) statistical fluctuations , ii ) the precision for the determination of the components of the four momentum vectors , iii ) contributions from any kind of background produced by means of the apparatus or due to physical reasons , iv ) the efficiency of the detector covering not 100 @xmath764 of the phase space due to size or construction limitations and finally v ) the reconstruction efficiency due to unavoidable inadequate performance whatsoever . here all collaborations have to meet the optimum between wishful thinking , technical possibilities and affordable costs . the high luminosity and 4@xmath9 detector celsius / wasa ( * w*ide * a*ngle * s*hower * a*pparatus ) facility @xcite is primarily designed for both measuring rare decay processes of light mesons and investigating meson production studies in the threshold region . the target and beam particles tend to be scattered at forward angles while the products of the meson decay will be distributed rather isotropically . the detector is constructed following these constrains and in full operation an annual production rate of about @xmath765 s is expected in the @xmath50 interaction . the essential parts of the celsius / wasa facility are : * the pellet target with : pellets diameter and frequency of 25 50 @xmath193 m and a few khz , respectively , + where the beam diameter is about 3 mm , @xmath766 atoms/@xmath752 . * the forward detector with : a segmented forward window counter ( fwc ) for fast timing in the first trigger level . + a forward proportional chamber ( fpc ) out of straw chambers to measure the scattering angle . + a forward trigger hodoscope ( fth ) built out of three layers of circular scintillators used for : + @xmath767fast trigger decisions , rough track position determination and particle identification . + a forward range hodoscope ( frh ) as a plastic scintillator calorimeter for measurements of + @xmath767particle energy and identification . + a forward range intermediate hodoscope ( fri ) adding a position sensitivity to the frh . * the zero degree spectrometer to filter out low rigidity reaction products employing hpge + or telescopes of si@xmath193-strip detectors , scintillation counters and csi(tl ) crystals . * the central detector with : a scintillator electromagnetic calorimeter ( sec ) consisting out of 1012 csi(na ) + @xmath767between the superconducting solenoid ( scs ) and the iron yoke . + a mini drift chamber ( mdc ) with 1738 straw tubes for measuring tracks of charged particles . the celsius / wasa detector is completed and regular data taking has started in late 2001 on some @xmath9production reactions where the @xmath44 s are detected by their @xmath768 decays . the tagging of the @xmath11production and the detection of the @xmath11 s by their @xmath768 and 3@xmath44 decays has been proven to work . in other words there is a high luminosity hadron facility ready for exciting physics . tof detection facility at the external beam line of cosy is a multi purpose modular arrangement of segmented barrel and circular planar scintillation detectors for measuring the time - of - flight and the energy loss of charged reaction products . in fig . [ tof_all ] two cad - drawings of a short and an extended version as well as a photograph of the outer barrel in it s longest assembly is shown . the modularity serves several purposes but essentially helps to adjust the system to the optimal dimensions according to the reaction under investigation . three scintillators placed 1 m , 0.5 m and 0.1 m in front of the target are used to veto a beam halo . further the following components are employed : * a few mm@xmath762 large liquid hydrogen / deuterium target with very thin windows @xcite * different granulated start counter systems optimized for the reaction type under investigation . * up to three barrels equiped with granulated cylindrical detectors for timing , spatial resolution and energy loss determinations . * granulated three layer stop counters , divided into inner and outer circular scintillation detector hodoscopes @xcite . * a scintillation calorimeter for energy determination of charged and neutral hadrons , placed behind the stop detector , see the short version of the facility , left in fig . [ tof_all ] . * in preparation a straw drift chamber hodoscope , to be installed behind the start detector systems . the cosy11 installation is an internal facility at cosy designed for threshold meson production studies . close to threshold , due to the small transvers momenta , all ejectiles are limited to a narrow forward cone in the laboratory system and therefore a high geometrical acceptance is attainable with a comparatively small detector set up . the essential components of the cosy11 installation are : * a cluster target located in front of a regular c shaped cosy dipole producing h@xmath754 or d@xmath754clusters of up to 10@xmath769 molecules and densities of about 10@xmath770 atoms/@xmath756 . * an exit window ( 187 @xmath771 7.6 @xmath752 ) of 30 @xmath193 m al and 2 @xmath771 150 @xmath193 m carbon fibers to separate the ultra high beam line vacuum from atmospheric pressure . it provides high acceptance compared to conventional magnetic spectrometers especially in the horizontal plane . * two sets of drift chambers stacks ( d1 and d2 in fig . [ cosy-11_detector ] ) with six and eight planes , respectively and a position resolution of about 200 @xmath193 m in each plane , for particle track reconstruction and momentum analysis through the magnetic field for positively charged ejectiles . * two scintillator hodoscopes ( s1 and s2 in fig . [ cosy-11_detector ] ) for fast level trigger determination . * a large area scintillation wall @xcite ( 2.2 @xmath771 1 m@xmath772 ) s3 in fig . [ cosy-11_detector ] placed 9.1 m downstream , for fast trigger , time of flight and consecutively invariant mass determination . * a detection system for negatively charged mesons consisting out of a scintillator detector and an array of silicon pad detectors , both mounted inside the dipole gap . * a monitor system for elastically scattered particles , utilized for luminosity determination . * in preparation for an extension : two wire chambers with a hexagonal drift structures to be positioned at the side and at the front exit windows and a threshold erenkov counter behind the s1 hodoscope for pion background subtraction . storage rings with the flexibility of different projectiles , variable beam momenta , extremely good phase space cooling and with the unrivaled options of internal and external detection systems provide an excellent tool for investigations on meson production at threshold . especially the high quality of relative and absolute energy resolutions available at modern accelerators are crucial for such studies . in the present report , discussing the close to threshold meson production in hadronic interactions , the natural restriction to investigations using only hadronic probes was made though interesting and informative results were certainly obtained with electromagnetic tools as well . the theory of strong interactions ( qcd ) has a convincing predictive power at high energies . perturbative expansions in the coupling constant no longer converge for medium energy scales and the structure of qcd becomes increasingly complicated . the threshold meson production in nucleon nucleon scattering offers valuable aspects for studying production , decay and interaction of hadrons and investigating the reaction mechanisms , the dynamics of the system and appropriately chosen subsystems as well as the structure of hadrons themselves . such experiments probe the sensitivity of nucleon nucleon interactions at distances of @xmath773 1 fm , while the long range part of the interaction is suppressed . meson production in @xmath774 reactions ( with @xmath775 ) is a process of strong inelasticity in proton proton scattering . in the present paper discussions utilizing polarized beam and/or targets are omitted . at the upgraded iucf cooler ring a polarized proton beam in combination with a windowless polarized hydrogen target now permits measurements of all spin correlation coefficients for @xmath776 reactions . very recently initial results for spin correlations were reported @xcite . the goal of such experiments is to qualify the energy dependence and the importance of higher partial waves by determining analyzing powers and spin correlations . here the hadron physics community certainly faces the challenge of new experimental techniques with necessary improvements of theoretical developments , since polarization observables are sensitive indicators of reaction mechanisms , relative contribution of partial waves and transition amplitudes . the production of @xmath11 mesons near threshold has been measured by different groups @xcite . very close to threshold only s waves are supposed to contribute to the production process and in fact the s wave production is large due to the presence of the n@xmath777(1535)s@xmath778 resonance which seems to act as an intermediate state . + both tof @xcite and cosy11 @xcite took recently rather high statistics data on the @xmath11 production at q @xmath779 15.5 mev with fairly flat angular distributions as expected for an s wave production . still whether some data suggest or do nt suggest differential cross section measurements might be too insensitive to reveal e.g. a small @xmath780contribution . data using a polarized beam , however , might indicate effects arising from an interference term of the kind : @xmath781 @xmath782 + where @xmath783 is the amplitude for the production of a final @xmath50 pair with angular momentum @xmath784 and @xmath785 while for @xmath786 the @xmath11 meson has an angular momentum of @xmath787 . + the importance of d waves in @xmath11 production is well known from both @xmath788 @xcite and @xmath789 @xcite reactions and the fact that this interference term is negative suggests that the @xmath11production in proton proton scattering is governed mainly by @xmath378exchange @xcite . it is the aim of the cosy11 group , followed by a recent feasibility study @xcite to investigate the presence of this and other higher partial waves through the study of the analyzing power @xmath790 . though the asymmetry should also be studied with respect to the final proton directions , the s d interference contributes significantly to the @xmath11 analyzing power defined as : @xmath791 @xmath782 + with the transition matrix element @xmath792 and the pauli matrix @xmath793 . + to lowest order in the @xmath11 d wave amplitude , the analyzing power is proportional to the imaginary part of the @xmath794@xmath56 interference @xcite , @xmath795 @xmath796 + where @xmath797 corresponds to the plane of the beam ( perpendicular to the polarization direction ) . on angular momentum grounds , this signal is expected to vary like @xmath798 , @xmath799 roughly like q. + the @xmath800 resonance could also be of importance for the @xmath11 production @xcite , leading to p waves in the final state . the interference between the @xmath243 and @xmath244 amplitudes gives an analyzing power proportional to : @xmath801 @xmath796 + and such a signal has been observed in the @xmath802 reaction at low energies @xcite . the presence of the three @xmath55waves in the interference might suppress such a term close to threshold for @xmath11 production and no unambiguous signal for them has been seen in the unpolarized differential cross section . nevertheless one must be prepared for their existence . the @xmath803@xmath780 interference gives a maximal a@xmath804 signal for @xmath805 and vanishes at @xmath806 , whereas the @xmath243@xmath244 signal should be largest at this point . + both the differential cross section and the analyzing power should be studied in order to detect effects associated with the onset of higher partial waves . it should be noted that the effects of incomplete acceptance are much less of a problem for @xmath790 than for the @xmath807 measurement provided that the beam characteristics of the two polarization states are similar . finally , the results should help to settle the on going discussion of whether the @xmath11 production is dominated by @xmath378 @xcite , @xmath390 @xcite or @xmath11 @xcite exchange . + first data on the @xmath11 meson production with a polarized proton beam in the reaction have been taken at the internal experiment facility cosy-11 the measurement was performed with a proton beam momentum @xmath808gev / c corresponding to an excess energy of @xmath809mev . + in figure [ ayvergleich ] a comparison of the data ( solid triangles ) with theoretical predictions for @xmath810mev from fldt and wilkin @xcite ( dotted line ) and nakayama @xcite ( solid line ) is shown . in addition , the theoretical calculations for @xmath811mev are plotted , both based on the one meson exchange model . while fldt and wilkin predict a dominance of @xmath378 exchange ( dashed line ) , nakayama concludes a dominant @xmath9 and @xmath11 exchange ( solid line ) . though the data are consistent with @xmath812 within the error bars given , there seems to be a slight increase of @xmath790 towards backward scattering angles . first results of the analyzing power deduced from the very first feasibility study run suffer from statistics but tend to differ from both theoretical calculations , although the predictions of @xcite show a quite good match . for a more precise statement further measurements are needed which could benefit from higher polarization now available at cosy . an important physics aspect of the modern cooler ring facilities is the interaction of particles in the final state . a detailed study of the dalitz plot of , e.g. the reaction @xmath813 measured at cosy-11 @xcite , made possible the extraction of valuable information about the @xmath518@xmath37 interaction . the @xmath814 reaction , measured at tsl in uppsala @xcite , limited significantly the range of values for the @xmath11@xmath37 scattering length this quantity is the relevant input for calculations studying the existence of @xmath11nucleus bound states . throughout the article the importance of an intermediate excitation of baryon resonances has been pointed out . + at excess energies of 12 gev , baryon resonances begin to play an important role in nucleon nucleon collisions . most of our understanding about these resonances comes from pion and photon induced reactions on the nucleon . current quark models predict many more resonances than can be experimentally extracted from the data . a large program at both jlab and elsa focuses on this so - called missing resonance problem and the search for yet unknown baryonic states in electro - nuclear reactions . however , besides the possibility that resonances thus far simply went unnoticed because the quality of the data was insufficient , another explanation might be that these resonances couple only weakly to photons or pions . an accelerator like cosy , built for nucleon nucleon as well as nucleon nucleus collisions will provide for the necessary experiments to fill that gap in the future . the investigation of symmetries and their breaking has always led to significant insights into the inner workings of nature . already a large program at cosy @xcite is studying charge - symmetry breaking - an effect closely linked to the difference in the current masses of the light quarks . + another interesting aspect of charge - symmetry breaking is its high sensitivity : once , for instance , the @xmath473@xmath482 mixing amplitude is isolated , it will provide us with insights about the nature of scalar resonances . studies about the structure of these resonances are under investigations at cosy @xcite . + at the celsius ring in uppsala the celsius / wasa detection system is in operation now and exciting experiments especially on @xmath11meson decay studies will start . naively , one would think that near - threshold meson production in pd collisions can be fully understood in terms of a short - ranged @xmath815 process , with the extra nucleon acting as a spectator . however , backward @xmath9 production in the @xmath816 can not be explained in terms of a spectator model . in the case of meson production in @xmath817 collisions the number of participating amplitudes is small as well , even though the transferred momentum is large . hereby a unique environment is available for studying the interaction of more than just two nucleons , e.g. three - nucleon forces . it is likely that eventually the research on three - nucleon forces will focus on this process . again , the role of polarization observables is important . for the experiments , the development of suitable experimental configurations , comprising polarized targets and associated detector systems that facilitate measurements of polarization observables is necessary . present and future experiments become possible only through the outstanding performance of the cooler accelerators . with the new injector for cosy the situation will be further improved . it will be possible to fill the machine up to the space - charge limit with polarized protons and deuterons , whereby electron - cooled beams at injection energy and stochastically - cooled beams at high energy of unprecedented brilliance will be available to the users . efficient correction schemes to preserve the polarization of protons during acceleration have already been established , while for deuterons no difficulties are anticipated , because no depolarizing resonances occur in the energy range of cosy . new polarization experiments in the future will certainly call for further refinement of the spin - manipulation systems in cosy , e.g. longitudinally polarized internal and external beams , for which suitable spin rotators have to be developed . the above mentioned examples which will be the subject of a summer school and workshop on cosy physics @xcite demonstrate the richness of physics in meson production reactions . the drawback of this wealth comes at the expense that , apart from rare cases , it is difficult to extract a particular piece of information from the data , e.g. both resonances and final state interactions modify the invariant mass plot in a coherent superposition . + fortunately , the use of polarization in the initial state acts as a spin filter and different contributions to the interaction can be singled out , because they show up in the angular distributions of different spin combinations . it is important to stress , that as long as only the lowest partial waves contribute significantly to the process under study , the number of independent observables from experiments with polarized beam and target is sufficient to allow for a model independent partial wave analysis . the authors would like to express their utmost graditude to the numerous people who gave the various help needed in order to successfully complete this article . + especially we would like to thank r. czyykiewicz , j. flood , j. haidenbauer , ch . hanhart , t. sefzick and p. winter for both scientific and technical support throughout the preparation of this manuscript . 99 -2pt v. bernard , n. kaiser and u.g . meiner , proceedings of the _ workshop on chiral dynamics : theory and experiment _ , july 2529 , 1994 , cambridge , massachusetts ; eds . a.m. bernstein and b.r . holstein ( springer , berlin ) , + c. wilkin , _ e print archive : _ nucl th/9810047 and + proceedings of the _ 8th international conference on the structure of baryons ( baryons 98 ) _ , september 2226 , 1998 , bonn , germany ; eds . d.w . menze and b. metsch + ( world scientific , singapore , 1999 ) , p.505 m. gell mann and k.m . watson , a.h . rosenfeld , d. koltun and a. reitan , m.e . schillaci et al . , h.o . meyer et al . , h.o . meyer et al . , w.w . daehnick et al . lee and d.o . riska , c. hanhart et al . , c. hanhart et al . , c. hanhart et al . , b. blankleider and a.n . kvinikhidze , s. wycech and a.m. green , a. fix and h. arenhvel , k.m . watson , m. batini , a. varc , t.s.h . lee , k. nakayama , _ e print archive _ nucl th/0108032 and + proceedings of the _ symp . on threshold meson production in pp and pd interactions _ , + june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 119 c. hanhart and k. nakayama , f. kleefeld , _ e print archive _ nucl th/0108064 and + proceedings of the _ symp . on threshold meson production in pp and pd interactions _ , + june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 51 c. hanhart , u. van kolck and g.a . miller , and references therein n.i . kochelev , v. vento and a.v . vinnikov , m. dillig , _ e print archive _ hep ph/0202067 n. kaiser , p.b . siegel and w.weise , n. kaiser , t. waas and w. weise , w. weise and u.g . meiner , hesr project , contributions to the scientific case , ( 1999 ) 103 d. harrach , proceedings of the _ 29th international workshop on gross properties of nuclei and nuclear excitations ( hirschegg 2001 : structure of hadrons ) _ , january 1420 , 2001 , hirschegg , austria ; eds . h. feldmeier , j. knoll , w. nrenberg and j. wambach ( gsi , darmstadt , 2001 ) r.l . workman , in r. machleidt and i. slaus , r. machleidt , a. moalem , e. gedalin , l. razdolskaja , and z. shorer , v. bernard , n. kaiser , u.g . meiner , e. gedalin , a. moalem , l. razdolskaja , e. byckling and k. kajantie , _ particle kinematics _ , ( john wiley & sons , new york , 1973 ) p. moskal et al . , meyer et al . , d. grzonka and k. kilian , proceedings of the _ symp . on threshold meson production in pp and pd interactions _ , june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 100 r.h . dalitz , j. dyring , _ ph.d . , university of uppsala , d. chodrow , d.a . hutcheon et al . , d.a . hutcheon et al . , m.g . bachman et al . , r. helmer , r. bilger et al . , h. caln et al . , h. caln et al . , h. caln et al . , p. moskal , _ e print archive : _ nucl ex/0110001 and + proceedings of the _ symp . on threshold meson production in pp and pd interactions + june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 27 r.j . slobodrian , m. lacombe et al . , s. hggstrm , _ ph.d . thesis _ , uppsala university , m. lacombe et al . , m. bscher et al . , cosy proposals no . 75 ( 1998 ) and no . 94 ( 2000 ) , + http://www.fzjuelich.de/ikp/anke f. duncan et al . , m. garon , j.w . van orden , _ e print archive : _ nucl th/0102049 , submitted to e. chiavassa et al . , r. smith and c. wilkin , j. smyrski et al . , f. hibou et al . , h. caln et al . , e. chiavassa et al . , a.m. bergdolt et al . , p. moskal et al . , _ e print archive : _ nucl ex/0110018 and j. bystricky and f. lehar , _ physics data _ nr.11 - 1 , fachinfo.zentrum , karlsruhe , 1987 , pt . i , p.98 h. rohdjess et al . , j. greiff et al . meyer and j.a . niskanen , p. moskal et al . , p. moskal et al . , _ e print archive : _ nucl ex/0201016 and p. senger et al . , j. schaffner bielich , i.n . mishustin and j. bondorf , s. goldhaber et al . , , + w. cameron et al . , adams et al . , c.j . adams et al . , , r. armenteros et al . , , + t.s . mast et al . , , m. sakitt et al , c. quentmeier et al . , f. balestra et al . , a. sibirtsev , w. cassing and c.m . ko , j. zomaczuk et al . , r. bilger et al . meyer et al . , j.d . jackson , j.m . blatt , h.a . bethe , d.y . wong and h.p . noyes , r.a . arndt , r.h . hackman and l.d . roper , e. oset , a. ramos , j. caro ramon , n. kaiser , s. wetzel , w. weise , a. bondar et al . , p. moskal et al . , m. wolke , c. quentmeier et al . , cosy proposal no . 61.2 ( 2000 ) , p. winter , m. siemaszko et al . , cosy proposal no . 61.3 ( 2001 ) , http://ikpe1101.ikp.kfajuelich.de m. wolke and a. sibirtsev , proceedings of the _ symp . on threshold meson production in pp and pd interactions , june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 239 r.a . arndt et al . , + the virginia tech partial wave analysis facility ( said ) http://gwdac.phys.gwu.edu h.o . meyer et al . , j.r . bergervoet et al . knutson , _ in nuclear physics at storage rings _ , edited by h.o . meyer and p. schwandt , + f. balestra et al . , m. abdel bary et al . , e - print archive : nucl - ex/0205016 , to be published in _ eur . h. caln et al . , a.b . migdal , b.j . morton et al . , naisse , h.p . noyes and h.m . lipinski , h.p . noyes , v.g.j . stoks et al . druzhinin , a.e . kudryavtsev and v.e . tarasov , r. shyam and u. mosel , m.l . goldberger and k.m . watson , _ collision theory _ , ( john wiley & sons , new york , 1964 ) s. stanislaus et al . , g. fldt and c. wilkin , p. moskal et al . , a. khoukaz et al . , _ annual report 2001 _ , institut fr kernphysik , universitt mnster , in press j.t . balewski et al . , s. sewerin et al . , p. kowina , _ ph.d . thesis _ , silesian university katowice ( 2002 ) , in preparation r. bilger et al . , s. marcello et al . , d. hesselbarth , _ ph.d . thesis _ , university of bonn , _ berichte des fz jlich _ , jl3849 ( 2001 ) w.j . fickinger et al . , g. fldt and c. wilkin , a. moalem et al . , , _ e print archive : _ hep ph/9505264 a.m. green and s. wycech , _ e print archive : _ nucl th/0009053 a. sibirtsev et al . , d. sigg et al . , , a.m. green and s. wycech , w. glckle et al . , u. schuberth , _ ph.d . thesis _ , university of uppsala , v. baru et al . niskanen , h.w . fearing , j. haidenbauer et al . , h. van haeringen , k. nakayama et al . groom et al . , m. batini et al . , _ e print archive : _ nucl th/9703023 j. smyrski , proceedings of the _ symp . on threshold meson production in pp and pd interactions _ , june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 46 a. fix and h. arenhvel , s.a . rakityansky et al . , a. svarc and s. ceci , contribution to n@xmath7772000 , _ e print archive : _ nucl th/0009024 c. wilkin , n.v . shevchenko et al . , s. wycech , h. garcilazo and m.t . pea , m. green , s. wycech , t. ueda , , t. ueda , s. wycech and a.m. green , a. deloff , f. hibou et al . , a. gillitzer et al . , cosy proposal no . 102 ( 2001 ) http://www.fzjuelich.de/ikp/cosytof g. rappenecker et al . tippens et al . , p. moskal et al . , _ annual report 2000 _ , ikp , fz jlich , jl3852 ( 2001 ) 48 p. moskal , _ ph.d . thesis _ , jagellonian university , cracow , + _ berichte des fz jlich _ , jl3685 ( 1999 ) http://ikpe1101.ikp.kfajuelich.de s. capstick et al . , _ e print archive : _ hep ph/0012238 and + _ hadronic physics white paper _ , proceedings of the _ aps division of nuclear physics town meeting on electromagnetic and hadronic physics _ , december 14 , 2000 , newport news , virginia ; + http://www.jlab.org/intralab/calendar/dnp@xmath818tm k. maltman and n. isgur , j.a . niskanen , j. haidenbauer , c. hanhart and j. speth , j.a niskanen , e. hernndez and e. oset , c.j . horowitz , h.o . meyer and d.k . griegel , h. machner and j. haidenbauer , j. adam et al . , t. sato et al . , t.d . cohen et al . , park et al . , u. van kolck , g.a . miller , d.o . riska , s.k . saha et al . , b. v. przewoski et al . , t. feldmann and p.kroll , _ e print archive : _ hep ph/0201044 and + proceedings of the _ workshop on eta physics : prospects of precision measurements with the celsius / wasa facility _ , october 2527 , 2001 , uppsala , sweden ; to be published in w. grein , p. kroll , m.m . nagels et al . , , k. nakayama , j. speth and t.s.h . lee , r. machleidt , j.f . germond and c. wilkin , j.m . laget , f. wellers , j.f . lecolley , g. fldt and c. wilkin , t. vetter , a. engel , t. bir and u. mosel , b.l . alvaredo and e. oset , m.t . pea , h. garcilazo and d.o . riska , p. winter , _ diploma thesis _ , university of bonn , + _ berichte des fz jlich _ , jl3943 ( 2002 ) , http://ikpe1101.ikp.kfa-juelich.de a. sibirtsev and w. cassing , r. pltzke et al . , p.ball , j.m . frre and m. tytgat , s.d . bass , s.d . bass , s. wetzel and w.weise , s.d . bass , _ e print archive : _ hep ph/0006348 and proceedings of the _ 8th int . workshop on + deep inelastic scattering and qcd ( dis 2000 ) _ , april 2530 , 2000 , liverpool , england ; + eds . j.a . gracey and t. greenshaw ( world scientific , singapore , 2001 ) , p.206 s.d . bass , n. nikolaev , published by the fz jlich in cooperation with canu . e. kou , s.d . bass , _ private communication _ s. d. bass , _ e print archive : _ hep ph/0108187 and + proceedings of the _ symp . on threshold meson production in pp and pd interactions _ , + june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 11 f. kleefeld , m. dillig and f. pilotto , m. dillig , _ e print archive : _ nucl th/0201050 s. okubo , , g. zweig , + j. iizuka , h. lipkin , h. lipkin , t.p . cheng and r.f . dashen , , t.p . cheng , , j.f . donoghue and c.r . nappi , , j. gasser , h. leutwyler , m.p . locher and m.e . sainio , , j. gasser , h. leutwyler and m.e . sainio , j. ashman et al . , , j. ashman et al . , , p. amaudruz et al . , t. adams et al e print archive : _ hep ex/9906038 and + proceedings of the _ workshop on physics with electron polarized ion collider ( epic 99 ) _ , april 811 , 1999 , bloomington , indiana ; eds . l.c . bland , j.t . londergan and a.p . szczepaniak ( world scientific , singapore , 2000 ) , p.337 v. barone , c. pascaud and f. zomer , and references therein j. ellis , m. karliner , d.e . kharzeev and m.g . sapozhnikov , t. gutsche , a. faessler , g.d . yen and s.n . yang , c.b . dover , t. gutsche , m. maruyama and a. faessler , l. bertolotto et al . , , + c. evangelista et al . , m.p . locher , y. lu and b.s . zou , m.p . locher and y. lu , v. mull , k. holinde and j. speth , d. buzatu and f.m . lev , , , + o. gortchakov , m.p . locher , v.e . markushin and s. von rotz , a.v . anisovich and e. klempt , u.g . meiner , v. mull , j. speth and j.w . van orden , v.e . markushin and m.p . locher , p. geiger and n. isgur , a.m. bertin et al . , f. balestra et al . , , v. blobel et al . , , r. baldi et al . , , m.w . arenton et al . , , + s.v . golovkin et al . , a. sibirtsev , a. sibirtsev and w. cassing , w.s . chung , g.q . li and c.m . titov , b. kmpfer and v.v . shklyar , a. sibirtsev and w. cassing , s. abd el samad et al . , m.p . rekalo , j. arvieux and e. tomasi gustafsson , k. nakayama et al . , k. nakayama et al . , g. hhler et al . , s. furuichi and k. watanabe , and 83 ( 1990 ) 565 r. machleidt , k. holinde and c. elster , p. mergell , u.g . meiner and d. drechsel , j. speth and r. tegen , c.b . dover and a. gal , s. capstick and w. roberts , n. kaiser , a.i . titov , b. kmpfer and b.l . reznik , f. hibou et al . , n. kaiser , v. flaminio et al . , v. herrmann , j. speth and k. nakayama , a.i . titov , b. kmpfer and b.l . reznik , d.o . riska and g.e . brown , j. ellis , m. karliner , d.e . kharzeev and m.g . sapozhnikov , r. jaffe , f. close , f. close et al . , d. morgan and m.r . pennington , f. kleefeld et al . , _ e print archive : _ hep ph/0109158 j. weinstein and n. isgur , z.s . wang , s. krewald and j. speth , k. hikasa et al . ( particle data group ) , d. lohse et al . , o. krehl , r. rapp and j. speth , j. haidenbauer , proc . of the _ symp . on threshold meson production in pp and pd interactions _ , june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 225 j.a . oller , e. oset and j.r . pelez , b.s . zou and d.v . bugg , n.a . trnqvist , r. kamiski , l. leniak and b. loiseau , r. kamiski , l. leniak and b. loiseau , r. kamiski , l. leniak and b. loiseau , j.a . oller , e. oset and a. ramos , d. alde et al . , , , + d. barberis et al . , d. barberis et al . , a. breakstone et al . , e.m . aitala et al . akhmetshin et al . , r.r . akhmetshin et al . , k. ackerstaff et al . , p. abreu et al . , k. ackerstaff et al . gay et al . , s. teige et al . , c. quentmeier , thesis _ , university of mnster , germany ( 2001 ) e.l . bratkovskaya et al . , v.y . grishina et al . , v. chernyshev et al . , cosy proposal no . 55 ( 1997 ) , m. bscher et al . , cosy proposal no . 55.2 ( 2001 ) ; http://www.fz-juelich.de/ikp/anke e. oset , j.a . oller , and u.g . meiner , p. moskal and w. oelert , proceedings of the _ symp . on threshold meson production in pp and pd interactions _ , june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 212 w. oelert , proposal for the cosy11 facility , cosy ( 1988 ) m. bscher et al . , cosy proposal no . 97 ( 2001 ) http://www.fzjuelich.de/ikp/anke s. steininger and u.g . meiner , u.g . meiner , _ e print archive : _ hep ph/0108133 , hep ph/0201078 and references therein m. alberg , j. aichelin and c.m . ko , , w. cassing , v. metag , u. mosel and k. niita , , u. mosel , , g.e . brown , c.m . wu and l.h . xia , , g.q . li , a. faessler and s.w . huang , , b.a . li , , g. hartnack , j. jaenicke , l. sehn , h.stcker and j. aichelin , , g.q . li and c.m . fang , c.m . li and y.m . zheng , m. lutz , a. ramos , and e.oset , d.b . kaplan and a.e . nelson , g.e . brown , c.h . lee , m. rho and v. thorsson , w. cassing and e.l . bratkovskaya , f. laue et al . , p. senger , and references therein h. oeschler , and references therein b. schrmann and w. zwermann , w. zwermann , j. randrup and c.m . ko , , g.q . li and c.m . ko , a. sibirtsev , k. tsushima , a. sibirtsev and a.w . thomas , g. fldt and c. wilkin , m. hoffmann et al . , a. deloff , m.m . nagels et al . , b. holzenkamp et al . , a. reuber et al . , g. alexander et al . , b. sechi zorn et al . , e. ferrari , , d.h . saxon et al . , , k.w . ball et al . , , + d.j . candlin et al . laget , k. tsushima , s.w . huang and a. faessler , , + g. giacometti et al . martin , p.m.m . maessen , t.a . rijken and j.j . de swart , k. tsushima , a. sibirtsev , a.w . thomas and g.q . li , r. shyam , f. balestra et al . vigdor , proceedings of the _ international workshop on flavor and spin in hadronic and electromagnetic interactions _ , september 2123 , 1992 , turin , italy ; eds . f. balestra , r. bertini , r. garfagnini ( ed . compositori , bologna , 1993 ) , p.317 a. bravar et al . , z. liang and c. boros , m. maggiora et al . , k. heller , proceedings of the _ 12th international symp . on high energy spin physics ( spin 96 ) _ , september 1014 , 1996 , amsterdam , netherlands ; eds . de jager , t.j . ketel , p.j . mulders , j.e.j . oberski , m. oskam tamboezer ( world scientific , singapore , 1997 ) , + s. paul , _ int . workshop on strange structure of the nucleon _ , march 1115 , 1997 , geneva , switzerland , j. felix , , a. sibirtsev , k. tsushima and a.w . thomas , a. sibirtsev , k. tsushima , w. cassing and a.w . thomas , n.g . kelkar , b.k . jain , r. siebert et al . tan , j.t . balewski et al . , a. gasparyan _ diploma thesis _ , moscow state engineering physics institute , 1999 a. gasparian et al . dalitz , j. mcginley , c. belyea and s. anthony , proceedings of the _ int . conf . on hypernuclear and kaon physics _ , 1982 , heidelberg , germany ; ed . b. povh , report mpi h82v20 ( max planck institut heidelberg , 1982 ) a. martin , j.j . de swart , d. cline , r. laumann and j. mapp , t.h . tan , a. kudryavtsev , r.h . dalitz and a. deloff , , e. oset et al . batty et al . , a.m. gasparian et al . , a. sibirtsev , k. tsushima , w. cassing and a.w . thomas , _ e print archive _ nucl th/0004022 r. shyam , g. penner and u. mosel , a. engel , r. shyam , u. mosel and a.k . majumder , j. dubach , w.m . kloet and r.r . silbar , l. tiator , c. bennhold and s.s . kamalov , j.m . laget , a. gasparyan , proc . of the _ symp . on threshold meson production in pp and pd interactions _ , june 2024 , 2001 , cracow , poland ; eds . p. moskal and m. wolke , + _ schriften des fz jlich matter and materials _ 11 ( 2002 ) 205 j. haidenbauer , w. melnitchouk and j. speth , proceedings of the _ 8th meeting on mesons and light nuclei _ , july 26 , 2001 , prague , czech republic ; j. haidenbauer , private communication a.t.m . aerts and c.b . dover , h. piekarz , o. braun et al . , f. hinterberger et al . , cosy proposal no . 92 ( 2000 ) j.a . gmez tejedor and e. oset , e. oset and m.j . vicente - vacas , l. alvarez ruso et al . , j. johanson , _ ph.d . , uppsala university ( 2000 ) w. brodowski et al . , d.r . f. cochran et al . cverna et al . , d.v . bugg et al . , l.g . dakhno et al . , f. shimizu et al . , d.c . brunt et al . , j. stepaniak , _ vi . int . oberjoch meeting on meson nuclear physics _ , oberjoch , germany ( 1996 ) l. alvarez ruso , _ ph.d . thesis _ , university of valencia ( 1999 ) a. abashian , n.e . booth and k.m . crowe , a. abashian et al . , j. banaigs et al . , g. grayer et al . , t. risser et al . , f. bellemann et al . , l. alvarez - ruso , ; + e. oset et al . , _ few body syst.suppl._ 11 ( 1999 ) 275 + l.alvarez-ruso , e.oset and e.hernandez , proceedings of the _ 7th international conference on mesons and light nuclei 98 _ , august 31 september 4 , 1998 , pruhonice , prague , czech republic ; eds . j. adam , p. bydzovsky , j. dobes , r. mach , j. mares , m. sotona ( world scientific , singapore , 1999 ) , p.81 , _ e print archive:_nucl - th/9810002 ; v. betker et al . , e. plouin et al . , c.l . hollas et al . , f. bonutti et al . , m.kermani et al . , f. bellemann et al . , _ annual report 2001 _ , ikp , berichte des fz jlich , 3978 , p44 j. ritman et al . t. waas et al . , ; t. waas , n. kaiser , and w. weise , ; t. waas , and w. weise , v. koch , a. baca , c. garcia recio , j. nieves , a. baldini et al . , _ total cross sections for reactions of high energy particles _ , ed . h. schopper , landolt brnstein , new series i/12 ( springer , berlin , 1988 ) j. berger et al . , b. mayer et al . , q. haider and l.c . liu , l.c . liu and q. haider , s. wycech et al . abaev and b.m.k . nefkens , h.c . chiang , e. oset and l.c . liu , m. betigeri et al . , r. bilger et al . , k. kilian and h. nann , g. fldt and c. wilkin , @xmath819 , _ the periodic table on the www _ , http://www.webelements.com , + i. mills et al . , _ quantities , units and symbols in physics chemistry _ , ( blackwell scientific publications , oxford , 1988 ) b. krusche et al . , , g.l . cheung and t.t.s . kuo , f. frascaria et al . , n. willis et al . , h.o . meyer , j.a . niskanen , m. betigeri et al . , p. hawranek , proceedings of the _ @xmath820 international conference on nuclear physics at storage rings _ , stori02 , june 1620 , 2002 , uppsala , sweden a. magiera and h. machner , a. magiera et al . , cosy proposal no . 59 , http://ikpgem02/gem / gem.html c. wilkin , j. bacelar et al . , cosy proposal no . 83.2 ( 2000 ) , http://www.fzjuelich.de/ikp/anke w.r . lozowski and j.d . hudson , v. jaeckle et al . , a. hassan et al . , a. khoukaz et al . , b. trostell , c. ekstrm et al . , c. ekstrm et al . , f. sperisen et al . , _ iucf sci . and techn . rep . _ 117 h. dombrowski et al . , and references therein h. caln et al . , m. dahmen et al . , g. anton et al . daehnick et al . , and references therein w. deinet et al . , c. wilkin , private communication t.s.h . lee , c. hanhart , private communication a.b . santra and b.k . jain , e. gedalin et al . , proceedings of the _ summer school and workshop on cosy physics _ , + august 30 september 4 , 2002 , kerkrade , the netherlands ; + organized by k. brinkmann , m. bscher , c. hanhart , a. khoukaz , a. lehrach , f. rathmann and m. wolke , http://www.fz-juelich.de/css2002 , to be published .

studies of meson production at threshold in the hadron hadron interaction began in the fifties when sufficient energies of accelerated protons were available . a strong interdependence between developments in accelerator physics , detector performance and theoretical understanding led to a unique vivid field of physics . early experiments performed with bubble chambers revealed already typical ingredients of threshold studies , which were superseded by more complete meson production investigations at the nucleon beam facilities triumf , lampf , psi , lear and saturne . currently , with the advent of the new cooler rings as iucf , celsius and cosy the field is entering a new domain of precision and the next step of further progress . the analysis of this new data in the short range limit permits a more fundamental consideration and a quantitative comparison of the production processes for different mesons in the few body final states . the interpretation of the data take advantage of the fact that production reactions close to threshold are characterized by only a few degrees of freedom between a well defined combination of initial and exit channels . deviations from predictions of phase space controlled one meson exchange models are indications of new and exciting physics . precision data on differential cross sections , isospin and spin observables partly but by no means adequately available are presently turning up on the horizon . there is work for the next years and excitement of the physics expected . here we try to give a brief and at the same time comprehensive overview of this field of hadronic threshold production studies . pacs numbers : + @xmath013.60.hb , 13.60.le , 13.75.cs , 13.75.-n , 13.85.lg , 13.85.ni , 13.85.rm , + @xmath014.40.cs , 25.40.-h , 25.40.ve , 29.20.dh

gravitational interactions involving `` hard '' binary stars ( i.e.those with orbital velocities that exceed the local velocity dispersion ) provide the energy source that drives globular cluster evolution from the time that the cluster core approaches collapse @xcite . these interactions have a profound effect on the long - term fates of both the cluster and the binaries . among the hardest binaries in clusters are compact , accretional x - ray sources . bright low - mass x - ray binaries ( lmxbs ) , which contain a neutron star primary , are strongly overabundant in dense clusters and thus almost certainly have a dynamical origin there ( see * ? ? ? * ; * ? ? ? * ) . there is also growing evidence that a majority of the more numerous cataclysmic variables ( cvs ) in clusters may also have a dynamical origin , rather than representing a primordial population @xcite . this is a complex issue , given the many possible channels for the formation of cvs in clusters , which include : the evolution of primordial binaries , mediated by hardening interactions ; exchange interactions between singles and primordial binaries , possibly involving multiple exchanges ; tidal captures which do not result in mergers ; and three - body interactions ( see * ? ? ? deep surveys of compact binary populations in globular clusters will help to clarify the relative importance of these cv formation mechanisms . ngc 6397 is the second closest globular cluster ( @xmath2 ; * ? ? ? * ) and is by far the nearest cluster to have undergone core collapse . it thus provides a special laboratory for studying the formation and evolution of compact binary systems in the extreme environment of a collapsed core . ngc 6397 has been a key target for _ rosat _ , hubble space telescope ( _ hst _ ) , and chandra x - ray observatory ( _ chandra _ ) studies of compact binaries in clusters . @xcite first detected cataclysmic variables in ngc 6397 , as counterparts to faint _ rosat _ sources , using _ hst _ wide - field / planetary camera 1 and @xmath0-band imaging to select -excess objects . @xcite confirmed the power of the photometric technique , providing _ _ faint object spectrograph verification of the first three cv identifications in ngc 6397 . @xcite carried out the first _ chandra _ observations of ngc 6397 , which detected a rich population of 25 sources within the cluster half - mass radius ( @xmath3 = 2.33 ; * ? ? ? wfpc2 imaging in , @xmath0 , @xmath4 , and @xmath5 has led to the identification of many of these sources , including nine cvs and a number of active binaries @xcite . since the study by @xcite was carried out , we have obtained considerably deeper _ chandra _ and _ hst _ imaging of ngc 6397 . the results of the _ chandra _ imaging , which detected 79 sources within the half - mass radius , are reported by @xcite . these x - ray sources , designated by a prefix u , are listed in table [ t : counterparts ] . in this paper , we report a search for optical counterparts to these sources using a deep _ hst _ acs / wfc dataset . we describe the data , the analysis method , and the results in the following sections . the optical imaging used in this study is the _ hst _ go-10257 dataset ( pi : anderson ) , which provides deep , highly dithered acs / wfc imaging of the central region of ngc 6397 in f435w ( @xmath1 ) , f625w ( @xmath0 ) , and f658n ( ) . one of the main motivations of this study was to search for wobbles in the positions of stars , indicative of massive unseen companions , such as white dwarfs , neutron stars , or black holes . to provide good time sampling , the center of ngc 6397 was imaged over 10 single - orbit epochs , spaced at approximately one - month intervals between 2004 july and 2005 june . a by product of this observing strategy is that we imaged the cluster at many different orientations , allowing for an exquisite mitigation of detector - dependent photometric and psf - related errors . in addition , the stacked images show round psfs with no sign of diffraction spikes . the dataset consists of 5 short @xmath1 ( 13s ) , 5 long @xmath1 ( 340s ) , 5 short @xmath0 ( 10s ) , 5 long r ( 340s ) , and 40 ( 390s ) exposures . the short exposures were designed to provide accurate photometry for stars above the main - sequence turnoff ( msto ) . saturation only becomes an issue near the tip of the giant branch . with the large number of frames , the psf sampling is particularly good for this band the _ hst _ data were analyzed using software based on the program developed for the acs globular cluster treasury project , described in @xcite . the routine first finds the astrometric and photometric mapping from each exposure into a master reference frame . it then searches through the master reference frame one small patch at a time , identifying stars where a significant number of exposures indicate coincident detections . these stars are then measured two ways : ( 1 ) they are measured in each individual exposure where they could be found , and ( 2 ) they are also measured simultaneously in all the exposures ( which provides a better measurement for the fainter stars ) . stars were measured with spatially variable , library psfs ( see anderson & king acs isr 2006 - 01 ) constructed from the go-9444 data set ( pi - cool ) of @xmath6cen , using an aperture of whole pixels that was optimized for the brightness of the star and its particular surroundings . since our best coverage was in , we did our star - finding on those images , and then used the positions found to identify and measure the stars in the b and r images . the final photometry we report here comes from averaging the photometry from the individual stars , with sigma clipping applied to remove outlying values due to cosmic rays , defective pixels , etc . a total of 25004 stars were detected . for the @xmath1 and @xmath0 bands , the photometry was performed independently for the short and long frames . photometric calibration to the vegamag system was adopted from @xcite . in addition to the star finding and measuring , we also produced stacked images of the scene , in a procedure that is akin to the iraf / stsdas _ drizzle _ algorithm with pixfrac set to zero . the stacked images were oversampled by a factor of two , in order to take advantage of the heavy dithering to increase the effective resolution . the resulting supersampled stacked images have a 12,000@xmath712,000 format and cover an approximately circular field of diameter 5 with a pixel scale of 0.025 . figures [ f : mosaic ] and [ f : mosaic_zoom ] show a combination of the 40 images , together with the error circles for the 79 _ chandra _ sources . the error circles are plotted at their actual sizes in both figures . as can be seen in these figures , the combination of images at various orientations results in a very smooth symmetric psf , with a distinct lack of diffraction spikes . color - magnitude diagrams ( cmds ) were constructed from the photometry using the @xmath0 magnitudes , and the and color indices . the short - exposure @xmath1 and @xmath0 magnitudes were used for stars brighter than about 2.5 mag below the msto , since the long - exposure frames saturate above this point . figures [ f : cmd_cv ] and [ f : cmd_ab ] show the resulting cmds . the efficacy of the photometric procedure is indicated by the tightness of the fiducial sequences . no proper - motion cleaning has been performed at this point in the analysis . proper - motion determination is described in [ astrometry ] . in [ source_id ] we will use proper motions in our source identification . unfortunately the quality of the cluster - field separation depends too sensitively on brightness to provide a definitive membership determination for all stars . the ( , @xmath0 ) cmd reaches deepest for the bluest stars , since the faintest red main - sequence ( ms ) stars are below the detection limit in @xmath1 . the upper part of the white dwarf ( wd ) cooling sequence is clearly detected in the ( , @xmath0 ) cmd , extending to nearly 10 mag below the msto in @xmath0 . there is also an indication of a second wd sequence above the primary sequence , which @xcite interpret as a he wd sequence . note that in the ( , @xmath0 ) cmd , the wds lie to the -deficit side of the ms , reflecting the strong absorption lines in wd spectra relative to those of faint ms stars . the blue stragglers and the blue horizontal branch stars similarly lie to the -deficit sides of the subgiant branch and the giant branch , respectively . the stacked _ hst _ acs / wfc images for each filter were rectified to a common astrometric system . we calculated an absolute astrometric zeropoint for the stacked images relative to the icrs using a two - step process . we first computed a plate solution for an eso wfi frame ( wfi.2002 - 06 - 21t01:45:08.153 ) using 357 primary astrometric standards from the usno ucac2 catalog . we then selected 350 secondary astrometric standards from the wfi frame and used these to compute a plate solution for our stacked acs / wfc images . we determined a boresight correction for the _ chandra _ source coordinates from @xcite by computing the mean offsets between the _ hst _ and _ chandra _ coordinates for the six brightest cvs and the millisecond pulsar ( msp ) psr j1740 - 5340 . this resulted in a shift of the _ chandra _ coordinates to the icrs of @xmath8 in ra and @xmath9 in dec , where the quoted uncertainties are the standard deviations of the offsets for the seven objects . as an external test of our boresight correction , we note that our optical position for psr j1740 - 5340 agrees with that of @xcite to @xmath10 in ra and @xmath11 in dec . they followed a similar astrometric procedure to what we used , with wfpc2 rather than acs / wfc imaging . we searched for optical counterparts to the _ chandra _ sources by overlaying the _ chandra _ error circles on the _ hst _ stacked images , with the boresight correction applied to the _ chandra _ source positions from @xcite . the 95% confidence error circle sizes were computed following @xcite . since the uncertainty in the optical positions @xmath12 ) was small compared with the size of the x - ray error circle radii ( @xmath13 ) , we neglected the contribution of the former to the total positional uncertainty . a preliminary proper - motion analysis was carried out in order to do a first - pass cluster - field separation . the present dataset was used for the reference epoch , with the stellar positions averaged over the 10 orbits . two different datasets were considered for the second epoch , go-7335 ( obtained with the wfpc2 in 1999 ) and go-10775 ( obtained with the acs / wfc in 2006 ) . the former provides a longer time baseline , while the later provides more depth and spatial coverage . to compute proper motions , we measured stars in each early - epoch exposure image , corrected these positions for distortion , then transformed these positions into the reference frame using linear transformations based on only cluster members . by using only members to define the transformations , we automatically get displacements ( and hence motions ) with respect to the cluster s systemic motion . the cluster - field separation that can be achieved for moderately bright stars is illustrated in fig . [ f : pm ] , which shows the proper - motion distribution for stars with @xmath14 . cluster and field regions are defined by a magnitude - dependent radius about the centroids of the apparent cluster and field clumps . stars that do not reside in either of these regions are considered to have an unknown proper - motion membership status . since the proper - motion error increases with magnitude , this proper - motion separation becomes less effective with increasing magnitude . eventually the cluster and field regions of the proper - motion plane show a strong overlap . it was generally possible to obtain an approximate discrimination between cluster and field stars to a limiting magnitude of @xmath15 . in a few cases where the 2006-second - epoch results were resulted in an unknown proper - motion membership status , the 1999-second - epoch results were definitive . optical cmds provide a powerful means of selecting likely _ chandra _ source counterparts and investigating their properties . for each of the 79 _ chandra _ sources within the half - mass radius , we checked the locations of all objects within the x - ray error circle in the cmds shown in figures [ f : cmd_cv ] and [ f : cmd_ab ] . objects that fell on the main sequence were considered to be unlikely counterparts , given the relatively low x - ray to optical flux ratio , @xmath16 , of most ms stars , in contrast to the ranges for chromospherically active binaries and cataclysmic variables . table [ t : counterparts ] summarizes the result of this counterpart search . plausible identifications were obtained for nearly all sources and are listed in this table , along with previous source designations , inferred object types , proper - motion information , and photometry . the only sources for which the error circle was empty are the quiescent low - mass x - ray binary ( qlmxb ) u24 @xcite , u104 , u106 , u111 , u113 , u119 , u121 , u122 , and u123 . in the case of u24 , there is a small `` blip '' near the center of the error circle in the stacked @xmath0 image , but it likely represents the combination of airy ring artifacts from two bright nearby stars . several other sources , u5 , u16 , u77 , and u84 , lie outside of the @xmath0 field and thus are difficult to classify , from these data , due to the lack of and color measurements . however , the u5 and u77 counterparts , v30 and v36 respectively , have been shown to be variable using ground - based photometry @xcite . @xcite classified v30 as an eclipsing binary that lies to the far red of the main sequence ( suggestive of it being a non - member ) . @xcite have recently classified v36 as a single - line spectropsopic binary . we have classified both of these stars as uncertain active binaries ( abs ) . u5 and u84 have very red @xmath17 colors , suggesting that their colors are likely very red as well , which supports an ab identification . table [ t : counterparts ] also gives a summary of the results of the proper - motion analysis , which indicates whether the object is consistent with being a cluster member , a field member , or neither , based on a comparison of its proper motion to the proper - motion distribution for stars of similar magnitude . based on the location of the proposed counterparts in the cmds ( or , in a few cases , based on ground - based detection of variability ) , we primarily assigned types of cataclysmic variable ( cv ) and chromospherically active binary ( ab ) . cvs were defined as being significantly to the blue of the ms and having significant excesses ( either relative to the ms or to the white dwarf sequence ) . abs were defined as lying within @xmath18 mag above the ms ( and thus within @xmath19 mag to the red of the ms ) and having small excesses ( @xmath20 mag ) , based on the previous analysis of abs in ngc 6397 by @xcite . in three cases where the only object in the error circle was an apparent ms star , u41 , u91 , and u112 , we note its presence in table [ t : counterparts ] . similarly , an apparent msto star is the only object present in the error circle for source u118 . we note that an ab with a low mass ratio and weak lines could look like a ms star in both cmds . two objects , u28 and u108 , were classified as background galaxies based on the extended appearance of their images . u28 resembles an edge - on spiral , while u108 has a more complex structure suggestive of interacting galaxies . the high x - ray to optical flux ratio of u28 ( see [ flux_ratio ] ) suggests that it is an active galactic nucleus ( agn ) . while the flux ratio of u108 was not determined , due to the complex nature of the optical image , it is clearly much lower than that of u28 . one possibility is that we are detecting x - ray emission from u108 that is produced by a galaxy collision . figure [ f : cmd_cv ] shows the location of the cv candidates in the cmds . there is a suggestion of an evolutionary sequence from young , bright cvs to old , faint ones . we return to this point in [ summary ] . the six brightest cvs mostly lie about 0.2 0.8 mag to the blue of the ms . the optical emission of the systems appears to be dominated by the secondary in the @xmath0 band , with a larger contribution from the disk in the @xmath1 band . the fairly high @xmath0-band flux indicates that the secondaries are relatively massive , @xmath21 , as inferred from the isochrones of @xcite . all of the bright cvs have substantial excesses relative to the ms ; these excesses generally increase with magnitude . the cv candidates 18 are numbered according to the scheme used in previous papers , as listed by @xcite . a prefix cv is used to designate each cv candidate . @xcite identify u28 as a cv ; however we find that this chandra source appears to be a galaxy . here we use the identification by @xcite of u60 as cv9 . cvs 1015 are newly identified in the present study ; @xcite assigned numbers for cvs 10 and 11 . we have numbered cvs 1215 based on optical luminosity within this group . the translation between _ chandra _ source numbers and cv numbers is given in table [ t : counterparts ] . there is a two magnitude gap between the bright cvs and the fainter cvs . just below this gap are two possible transitional objects , u7 and u61 , which lie between the white dwarf sequence and the main sequence . we note that the proper - motion data produce an unknown membership status for u61 , while u7 registers as consistent with membership . this leaves open the possibility that u61 is a background agn with an emission line that is shifted into the window . however , the image of the u61 counterpart does not show obvious evidence of extension . the fainter cv candidates , below u7 and u61 , lie in the vicinity of the wd cooling sequence . the optical fluxes for the faint cv candidates are clearly dominated by the contribution of the wd . the indices for these faint cvs nearly all lie at least 0.5 mag to the excess side of the wd clump , which is centered at about @xmath22 . this suggests that the faint cvs have a strong -emission core ( due to an accretion disk ) within the broad absorption lines of the wd continuum . this inferred spectroscopic property of the faint cvs appears to be generally consistent with those of the wz sge class of evolved cataclysmic variable @xcite . another possibility is that some of these objects are am cvn type double - degenerate systems , in which a low - mass he wd donor feeds a much more massive carbon - oxygen wd . such systems show no evidence of hydrogen lines in their spectra and instead have he i and sometimes he ii lines , which are typically in emission @xcite . am cvn stars could have evolved into contact from the known he wds in ngc 6397 , which have been shown to have ( detached ) heavy co wd companions @xcite . such an object should lie close to the main sequence in the @xmath23 cmd . possible candidate am cvn stars among the cv candidates reported here are u25 , u31 , u60 , u80 , and u83 . we note that most of the wd - like cv candidates cluster near a magnitude of @xmath24 , which corresponds to a @xmath25 , using the distance modulus determined by @xcite . we note that @xcite have reported a sharp peak ( `` spike '' ) near the period minimum ( @xmath26 ) in the period distribution of cvs selected from the sloan digital sky survey ( sdss ) . the average absolute magnitude for the cvs in the spike is @xmath27 . allowing for the difference in photometric band , this is reasonably close to the characteristic magnitude of the faint cvs in our study . this suggests that these faint cvs may well belong to the spike population , i.e. have periods in the vicinity of the period minimum . this could be tested by determining orbital periods for the faint cvs from suitable time - resolved photometry . @xcite find that of 33 cvs with orbital periods less than 86 min , 20 have spectra that are wd - dominated . in our sample , it appears that all seven faint cvs are wd - dominated . given the small sample size , it is not clear whether this represents a significant difference between faint field and faint cluster cvs . figure [ f : cmd_ab ] shows the location of the objects identified as ab stars in the cmds . we note that these objects sometimes lie near the edge of the ms in either the right or left panel , but deviate by a larger amount in the other panel . of the 42 ab counterparts listed in table [ t : counterparts ] , 26 are newly identified in our study . overall , the abs appear to form a relatively homogeneous binary sequence alongside the ms , presumably mostly differing in mass . @xcite have argued that these stars are likely by draconis stars and represent a hard binary population in ngc 6397 . in a large study of photometric variability in 47 tuc , @xcite found a substantial population of by draconis binaries , as well as w uma and other contact binaries , and eclipsing binaries . in addition to the likely by dra population , there are a number of counterparts that show a significantly different distribution in the cmd . these include the `` red straggler '' counterpart to u12 ( the millisecond pulsar ) and the similar counterpart to u18 , the blue - straggler counterpart to u70 , a group of five stars with very red colors , and a group of five stars with strong excesses . there are two stars in common between these latter two groups . the counterparts to u12 and u18 both have proper motions consistent with cluster membership . given their similar optical and x - ray properties , it appears likely that u18 is a msp @xcite . the counterpart to u70 , which registers as a proper - motion cluster member , has been identified as variable star v20 by @xcite . @xcite find that it is a double - line spectroscopic binary with a period of @xmath28 and a mass ratio of @xmath29 . the he wd candidate pc-5 @xcite , which has been previously suggested as a possible counterpart to u70 , lies well outside of the current error circle . the counterparts to u42 , u63 , u65 , u86 , and u92 differ strikingly from the typical abs , lying up to 1 mag to the red of the ms , and thus well above it . of these , the counterparts to u42 and u92 have been previously identified as optically variable stars @xcite , as confirmed here ( see [ variability ] and fig . [ f : variability ] ) ; the counterparts to u63 , u65 , and u86 are new identifications . the u92 counterpart ( v7 ) is classified as a w uma system by @xcite , while the u42 counterpart ( v26 ) is classified as an irregular variable by @xcite . u92 is consistent with the cluster proper - motion distribution . the proper - motion information available for u42 , u65 , and u86 produce an unknown membership status . these stars have proper motions that are well outside of the cluster distribution but they are also inconsistent with the field distribution . there is no proper - motion information for u63 . one reasonable possibility is that these stars are foreground active binaries superposed on the cluster , as discussed by @xcite in the case of u5 ( v30 ) . in this interpretation , the foreground binaries contain k / m type dwarfs with distance moduli that put them well above the cluster main sequence in the @xmath30 cmd . figure [ f : cmd_ab ] also shows several ab candidates that have larger than typical excesses ; these include the counterparts to u62 , u65 , u66 , u86 , and u120 . there is no proper - motion information available for u62 . the proper - motion information for u65 , u86 , and u120 produces an unknown membership status . u66 is consistent with the field proper - motion distribution . thus , the proper - motion information does not offer much assistance in determining whether these stars are likely cluster members . @xcite have drawn attention to two of the variable stars that we have identified as ab counterparts , v17 and v36 , which correspond to the sources u76 and u77 . these similar objects both lie close to the main - sequence turnoff in the cmd presented by @xcite . both objects are single - line spectroscopic binaries , in which the unseen primary is inferred to have a mass in excess of 1 . they suggest that these two objects are members of a class of dormant degenerate binaries which may have much larger peak x - ray luminosities . as an additional aid to the classification of sources , we have examined the x - ray to optical flux ratio , @xmath16 , where we take @xmath31 from @xcite and set @xmath32 . the latter conversion factor is computed from the _ hst _ flux calibration constants . the x - ray count to flux transformation assumes a power - law spectrum with a photon index of @xmath33 @xcite . the resulting flux ratio is plotted versus @xmath34 in fig . [ f : fx_fopt ] . the ratio is observed to be higher for accretional sources , such as cvs , lmxbs , and agns , than for chromospherically active binaries ( see e.g. * ? ? ? * who studied the _ chandra _ source distribution in m4 ) . consistent with this , fig . [ f : fx_fopt ] shows that the cv and ab candidates form two distinct groups with no overlap in flux ratio . the median flux ratio is about 500 times larger for the cvs than for the abs . within each group there is a broad range of the flux ratio , about a factor of 40 for the cvs and a factor of 100 for the abs . there is no apparent dependence of the flux ratio on @xmath34 over a range of more than three decades in @xmath34 for the cvs and nearly two decades for the abs . the apparent agn , u28 , has the highest flux ratio of all objects . the known millisecond pulsar , u12 , has a higher flux ratio than most of the abs and is about 40@xmath7 brighter in the x - ray than a typical ab . the object u18 , which has a similar optical counterpart to that of u12 , has even higher values of @xmath34 and @xmath16 . this supports the interpretation of u18 as a msp not yet detected at radio wavelengths @xcite . interestingly , the x - ray flux of cv5 ( u22 ) is considerably lower in the 2007 observations than in the previous 2002 observations ; the flux differs more than an order of magnitude . this gives cv5 the lowest @xmath16 value of all the cvs when the 2007 x - ray flux is used together with the 2004 - 05 optical flux . since the x - ray and optical fluxes were not measured contemporaneously , it is possible that cv5 typically has a flux ratio that is more consistent with the other bright cvs . we note that the x - ray flux of a bright cv typically drops during outburst ( e.g. * ? ? ? * ) , suggesting that cv5 may have been in outburst in 2007 . the objects u62 , u66 , and u86 , which lie among the abs with the four highest flux ratios , all have significantly larger excesses than the typical ab . it is not surprising that two measures of chromospheric activity , x - ray flux and flux , are correlated for these stars . since our dataset provides a 40-exposure time sequence of exposures , with four exposures per orbit , it was possible to investigate optical variability . the time sequence samples time scales shorter than about one hour ( the visibility period per _ hst _ orbit ) and also time scales from one month to one year . the analysis of variability is complicated by the presence of outliers in the time sequences , some of which represent photometric anomalies . we investigated several measures of variability and adopted the rms deviation about the mean of the magnitude measurements , computed using an iterative 3@xmath35 clip to reduce the impact of outliers . we plot @xmath36 versus mean magnitude in figure [ f : variability ] . this procedure results in a measure of variability that is most sensitive to orbital variability of binary systems , rather than large - amplitude fluctuations of cvs , given the outlier filtering . thus , we also investigated the total range of the magnitudes for each object and constructed light curves for each of the cv candidates in order to test for outburst behavior . as can be seen in fig . [ f : variability ] , most stars fall on a well - defined sequence of increasing @xmath36 with increasing magnitude . stars that lie significantly above this sequence generally show evidence for variability , although some of this scatter is likely due to photometric issues . we note that the mean trend of the variability index @xmath36 with mean @xmath37 magnitude differs for stars brighter than @xmath38 , for which @xmath36 increases from a minimum of about 0.003 mag to about 0.04 mag for the brightest stars plotted . this behavior for bright stars is due to saturation issues that increase in size with decreasing magnitude . we have plotted the positions of the previous and new chandra source counterparts in fig . [ f : variability ] . many of these counterparts show significant variability as measured by the @xmath35 of the time series . this group includes all of the bright cvs 16 : sources u10 , u17 , u19 , u21 , u22 , and u23 . of the newly identified faint cv counterparts , only u61 registers as clearly variable by this measure and u25 shows a possible slight variability . while the other faint cv candidates have @xmath36 values that are comparable in size to those of the bright cvs , these values are not significantly larger than those of other stars of similar magnitude . thus , the failure to detect statistically significant variability for the other faint cv counterparts may well be due to the decreasing sensitivity of the variability test with increasing magnitude . the total magnitude range for each of the six bright cvs was about 0.61.3 mag , which is less than the 25 mag amplitude for a dwarf nova outburst @xcite . source u21 ( cv4 ) had the largest amplitude ( 1.3 mag ) of these bright cvs , with a light curve that a high phase of about 50 days duration . source u61 ( cv12 ) displayed a total range of 2.3 mag . the faintest magnitude measurements , which contribute significantly to this large range , have a high degree of uncertainty . the light curve shows a clearly differentiated low state in the first half of the year of observation and a high state during the second half . the mean offset between these two states is about 1 mag and thus also falls short of a dwarf nova outburst . a number of the candidate ab counterparts also show evidence of variability , including six newly identified abs : sources u79 , u97 , and u114 , which show strong variability at the level of the six bright cvs ; sources u88 and u101 , which show moderate variability ; and source u103 , which shows slight evidence of variability . five previously identified ab counterparts , u14 , u18 , u42 , u70 , and u92 , show clear variability , and u76 and u81 show weaker evidence of variability . we note that u18 has a similar level of variability to the msp counterpart u12 , again suggesting that these are similar objects . we carried out a period - folding light curve analysis for the bright cvs 16 and the msp u12 . given the very unequal spacing of the data , it is difficult to determine a period ab initio , due to the high degree of aliasing . however , we were able to confirm the previously determined orbital periods obtained by @xcite and @xcite for cvs 1 and 6 , as well as that of u12 @xcite . we also see evidence for additional long - term variability for cvs 1 and 6 , at the level of a few tenths of a magnitude . we determined the cluster center by iterative centroiding in a 20 radius aperture using a sample of ms stars with magnitudes in the range @xmath39 . the resulting center of @xmath40 lies within about 1 of cv2 ( u19 ) . experimentation with the centroiding aperture size and the stellar sample definition indicates that the center position is uncertain by about 1 . we then determined the cumulative radial distributions of a number of stellar groups out to a radius of 100 , which is the approximate size of the central surface density cusp of ngc 6397 @xcite . figure [ f : radial ] shows the distribution functions for the stellar groups listed in the figure legend . as has been previously noted , the bright cvs in ngc 6397 show a strong central concentration relative to other stellar groups , including the fainter cvs . with the exception of cv6 , the five other bright cvs lie within 11 ( @xmath41 ) of the cluster center . in comparison , with the exception of u25 , the faint cvs lie at least 20 ( @xmath42 ) from the cluster center . we carried out kolmogorov - smirnov intercomparisons of the stellar samples defined in fig . [ f : radial ] . the comparison of each group to a msto sample , defined by @xmath43 , is given in table [ t : cored_pl_fits ] , where the probability , @xmath44 , of the two samples being drawn from the same parent distribution is listed . the bright cv and blue straggler ( bs ) samples both differ very significantly from the msto sample ( @xmath45 ) , while the overall cv and ab samples differ from the msto sample at the 4% level and 1% level , respectively . the bright and faint cv samples differ from each other at a significance level of 3% and the bright cv and ab samples differ at a level of 2% . lrcccccr msto & 1111 & 1.0 & @xmath46 & @xmath47 & @xmath48 & & + cv & 14 & @xmath49 & @xmath50 & @xmath51 & @xmath52 & 2.3 & 4.0% + bright cv & 6 & @xmath53 & @xmath54 & @xmath55 & @xmath56 & 2.9 & 0.07% + faint cv & 8 & @xmath57 & @xmath58 & @xmath59 & @xmath60 & 0.2 & 89% + ab & 36 & @xmath61 & @xmath62 & @xmath63 & @xmath64 & 2.6 & 1.1% + bs & 23 & @xmath65 & @xmath66 & @xmath67 & @xmath68 & 3.7 & 0.03% + in order to further investigate the spatial distribution of the _ chandra _ sources in ngc 6397 and the implications for object masses , we carried out maximum likelihood fits of what has been termed a `` generalized king model '' to the surface density distributions . this model can be described more generally as a `` cored power law , '' in that it has a smooth transition from an inner core region to an outer power law . it takes the form , @xmath69^{\alpha/2},\ ] ] with the core radius @xmath70 related to the scale parameter @xmath71 by , @xmath72 an `` analytic king model , '' which provides a good fit to the surface density profile of the turnoff - mass stellar group for a normal - core cluster ( e.g. 47 tuc ; * ? ? ? * ) , corresponds to @xmath73 and @xmath74 . since ngc 6397 has a central - cusp structure , due to its core - collapsed status , we expect that the turnoff - mass group will have a flatter power - law slope than @xmath75 @xcite . we performed a maximum - likelihood fit of eqn . [ eqn : cored_pl ] to a turnoff - mass group , defined by @xmath43 and radial offset within @xmath76 of the cluster center , to determine the best - fit values of @xmath77 and @xmath70 . we used nonlinear optimization to maximize the likelihood and bootstrap resampling to estimate the parameter uncertainties , as discussed by @xcite . the resulting best - fit parameter values are @xmath78 and @xmath79 . by fitting eqn . [ eqn : cored_pl ] to the surface density distributions of individual groups such as the cvs and abs , the characteristic masses of these objects can be estimated , as described by @xcite and @xcite . in order to do this , we adopt the approximation here that stellar groups with masses exceeding the turnoff mass are in approximate thermal equilibrium with the turnoff mass group and that the velocity dispersion of each mass group is approximately constant with radius within the central cusp region . this approach is motivated by fokker - planck simulations of post - collapse core oscillations which indicate that a cluster core spends a majority of the time near a maximally expanded state in which the more massive stellar groups are close to thermal equilibrium @xcite . more recent monte - carlo and n - body simulations suggest that the presence of primordial binaries may delay the gravothermal oscillation phase to well beyond a hubble time and that even clusters with an observationally core - collapsed designation such as ngc 6397 may still be in the binary burning phase @xcite . in any case , the assumption of thermal equilibrium among the most massive components in the central region of an evolved cluster is a useful starting point . in the thermal equilibrium approximation , the surface density profile for a mass group with mass @xmath80 is given by eqn . [ eqn : cored_pl ] with a slope parameter @xmath77 related to the turnoff - mass slope @xmath81 by @xmath82 where @xmath83 . thus , fitting this model involves just one free parameter , the mass ratio @xmath84 , as @xmath85 is determined by normalization and @xmath71 and @xmath81 are determined by the fit to the turnoff - mass group . table [ t : cored_pl_fits ] gives the results of maximum - likelihood fits of eqn . [ eqn : cored_pl ] to the turnoff mass stars , cvs , abs , and a comparison group of blue stragglers ( bss ) . the cumulative radial distributions for these groups and the model fits are shown in fig . [ f : radial ] . as can be seen from the table , the @xmath84 values for all of the groups exceed unity , indicating that the characteristic masses exceed the turnoff mass . for all cases other than the faint cvs , the excesses are significant at the 2@xmath35 or higher level . the groups with the strongest evidence for a mass significantly above the turnoff mass are the bright cvs ( @xmath86 ) and the bss ( @xmath87 ) . we note that the results of this analysis agree with the k - s comparison results given in the last column of the table . the inferred mass range for the bright cvs ( @xmath88 ) is consistent with a system made up of a heavy white dwarf ( e.g. @xmath89 ) with a main sequence secondary of mass @xmath90 . the inferred mass range for the faint cvs ( @xmath91 ) is consistent with a somewhat less massive white dwarf ( e.g @xmath92 ) and a secondary mass that has been whittled down to @xmath93 . in such a system , the optical flux would be dominated by the white dwarf , as observed here . we presume that the white dwarf in this case has been heated by the accretion process , since noninteracting heavy white dwarfs are too faint to detect due to the degree of cooling that has occurred over their long lifetimes . as discussed in [ source_types ] , an alternative interpretation is that some of the cvs are am cvn systems , in which the secondary is a he wd . in this case , the mass of the secondary would likely be about 0.20.3 @xcite . it is interesting to compare the white dwarf masses inferred here with those for field cvs . @xcite have recently determined a mass of @xmath94 for the wd in a field cv that lies above the period gap . @xcite have determined the masses of the wds in seven cvs that lie below period gap , with a resulting mean mass and standard deviation of @xmath95 . thus , given the uncertainty ranges in the cv masses inferred here , our results for likely wd masses are generally consistent with these measured wd masses . we have shown that with moderately deep @xmath1 , @xmath0 , and _ hst_-acs / wfc imaging , it is possible to identify and classify nearly all of the 79 currently detected _ chandra _ sources within the half - mass radius of ngc 6397 . our classifications are based on cmd location and are tested for consistency with expected ranges of the x - ray to optical flux ratio . in only a few cases was it not possible to classify the source , generally due to a lack of coverage in one filter . the vast majority of the sources in the cluster can be classified as either cataclysmic variables ( a total of 15 ) or active binaries ( a total of 42 ) . the x - ray to optical flux ratio for the cvs ( 0.2 6 ) substantially exceeds that for most of the abs ( @xmath96 ) . there is one previously known msp and one potential msp ; the latter classification is based on the similar photometric properties of the two objects . the proper - motion test for cluster membership produces an unknown status for many of the _ chandra _ source counterparts . this is largely a consequence of the short time baseline of 1.5 years between the two epochs of acs / wfc data used for determining proper motions . while the 1999 wfpc2 observations give a longer baseline , fewer stars are detected by the wfpc2 . thus , the proper - motion part of this study would benefit substantially from new _ hst _ observations of ngc 6397 with either the acs / wfc or the wfc3 . while it appears likely that the vast majority of the proposed counterparts are cluster members , proper - motion confirmation would be valuable . it would be particularly interesting to know how many of the atypical abs are cluster members . while it is likely that most are foreground stars , one of the very red ones has appears to be a member on the basis of our preliminary proper - motion measurements . the cv distribution in ngc 6397 has a bimodal character . we detect distinct bright and faint populations with only two possible transition objects . the optical emission of the six brightest cvs appears to be dominated by a relatively massive secondary , while that of the faint cvs appears to be dominated by the wd , with very little contribution from a very low - mass secondary . alternatively , some of the faint cvs may be am cvn systems , with a he wd secondary . the distribution of cvs in optical luminosity and color is consistent with expected cv evolution due to the effects of mass transfer and angular momentum loss . as cvs age and the secondary loses mass to the primary , the orbit tightens due to magnetic braking and ultimately gravitational wave radiation . in this picture , the bright cvs 16 represent young , recently formed systems , while the faint cvs represent old , highly evolved systems . while cvs can be born at any orbital period , depending on the nature of the secondary star , there is a bias for cvs formed in exchange interactions in collapsed cluster cores to have more massive secondaries and thus longer orbital periods than newly formed field cvs . this is both the result of mass segregation in the cluster , which may invert the mass function in the cluster core , and of the tendency for more massive stars to displace less massive ones in exchange interactions . thus , while it is possible for a faint cv to be a young system , we expect that the bulk of the faint cvs evolved from an earlier bright state . given the likely ongoing formation of cvs by dynamical processes in the central region of the cluster , it is plausible that we are observing a roughly equilibrium cv population in ngc 6397 . however , there remains a possibility that the current population of six bright cvs represents the result of a recent core collapse event that resulted in an enhanced production of cvs . in this case , the bright cvs would be overpopulous relative to the faint cvs . the six brightest cvs in ngc 6397 have a much more centrally concentrated spatial distribution than do either the fainter cvs or the abs . indeed , five of six bright cvs are located within about 2@xmath70 of the cluster center , while all but one of the fainter cvs lie outside of 4@xmath70 . this suggests that cvs are preferentially formed by dynamical processes within the core and surrounding region and migrate to larger radii as they age and undergo repeated scattering interactions that act to increase the typical size of their orbits in the cluster potential . since repeated strong interactions between compact binaries and other stars ultimately result in the ejection of the binaries from clusters , the lifetime of a compact binary in a cluster is extended if the amount of time it spends in the central region of the cluster is reduced . thus , it is reasonable to expect that long - lived cvs in globular clusters should generally be found outside of the densest regions . this is consistent with the spatial distribution of the cvs in ngc 6397 . , s. f. , becker , a. c. , haggard , d. , prieto , j. l. , knapp , g. r. , sako , m. , halford , k. e. , jha , s. , martin , b. , holtzman , j. , frieman , j. a. , garnavich , p. m. , hayward , s. , ivezi , . , mukadam , a. s. , sesar , b. , szkody , p. , malanushenko , v. , richmond , m. w. , schneider , d. p. , & york , d. g. 2008 , , 135 , 2108 , s. f. , haggard , d. , homer , l. , joshi , n. r. , margon , b. , silvestri , n. m. , szkody , p. , wolfe , m. a. , agol , e. , becker , a. c. , henden , a. , hall , p. b. , knapp , g. r. , richmond , m. w. , schneider , d. p. , stinson , g. , barentine , j. c. , brewington , h. j. , brinkmann , j. , harvanek , m. , kleinman , s. j. , krzesinski , j. , long , d. , neilsen , jr . , e. h. , nitta , a. , & snedden , s. a. 2005 , , 130 , 2230 , b. t. , dillon , m. , southworth , j. , thorstensen , j. r. , rodrguez - gil , p. , aungwerojwit , a. , marsh , t. r. , szkody , p. , barros , s. c. c. , casares , j. , de martino , d. , groot , p. j. , hakala , p. , kolb , u. , littlefair , s. p. , martnez - pais , i. g. , nelemans , g. , & schreiber , m. r. 2009 , , 397 , 2170 , d. , & hut , p. 2003 , the gravitational million - body problem : a multidisciplinary approach to star cluster dynamics ( the gravitational million - body problem : a multidisciplinary approach to star cluster dynamics , by douglas heggie and piet hut . cambridge university press , 2003 , 372 pp . ) , d. , lewin , w. h. g. , anderson , s. f. , baumgardt , h. , filippenko , a. v. , gaensler , b. m. , homer , l. , hut , p. , kaspi , v. m. , makino , j. , margon , b. , mcmillan , s. , portegies zwart , s. , van der klis , m. , & verbunt , f. 2003 , , 591 , l131 llrlllllcccl u5 & 17:40:54.531 @xmath9753:40:44.57 & 1.85 & v30 & 1 & ab ? & ? & & & 18.73 & 19.23 & outside of @xmath0 field + u7 & 17:40:52.832 @xmath9753:41:21.77 & 1.81 & cv10@xmath98 & 1 & cv & c & & 22.58 & 23.44 & 24.73 & + u10 & 17:40:48.978 @xmath9753:39:48.62 & 1.21 & cv6 , v13 & 1 & cv & c & c & 19.14 & 20.34 & 21.53 & + u11 & 17:40:45.781 @xmath9753:40:41.52 & 0.58 & cv7 & 1 & cv & ? & c & 23.65 & 24.26 & 25.88 & + u12 & 17:40:44.621 @xmath9753:40:41.60 & 0.42 & by - wf4 - 1 , v16 & 1 & msp & c & c & 16.25 & 17.43 & 18.99 & + u13 & 17:40:44.084 @xmath9753:40:39.17 & 0.33 & cv8 & 1 & cv & ? & c & 24.00 & 24.33 & 25.95 & + u14 & 17:40:43.328 @xmath9753:41:55.46 & 1.46 & v35 & 1 & ab & c & & 18.31 & 19.72 & 20.96 & + u15 & 17:40:42.910 @xmath9753:40:33.81 & 0.14 & by - pc-2 & 1 & ab & c & & 17.25 & 18.53 & 19.93 & + u16 & 17:40:42.636 @xmath9753:42:15.24 & 1.78 & & 1 & ? & ? & & & & 25.83 & outside of @xmath0 and @xmath1 fields + u17 & 17:40:42.651 @xmath9753:40:19.30 & 0.17 & cv3 , v33 & 1 & cv & c & c & 18.52 & 19.10 & 20.67 & + u18 & 17:40:42.606 @xmath9753:40:27.62 & 0.07 & by - pc-8 , v31 & 1 & msp & c & c & 16.10 & 17.51 & 18.73 & + u19 & 17:40:42.306 @xmath9753:40:28.70 & 0.02 & cv2 , v34 & 1 & cv & c & c & 18.87 & 20.02 & 21.03 & + u21 & 17:40:41.830 @xmath9753:40:21.37 & 0.13 & cv4 & 1 & cv & c & c & 19.82 & 20.94 & 21.82 & + u22 & 17:40:41.701 @xmath9753:40:29.00 & 0.07 & cv5 & 1 & cv & c & c & 20.15 & 21.50 & 22.04 & + u23 & 17:40:41.597 @xmath9753:40:19.30 & 0.18 & cv1 , v12 & 1 & cv & c & c & 17.88 & 18.93 & 20.25 & + u24 & 17:40:41.468 @xmath9753:40:04.43 & 0.42 & & 0 ? & qlmxb & & & & & & hint of detection in @xmath0 only + u25 & 17:40:41.237 @xmath9753:40:25.79 & 0.15 & cv13@xmath98 & 1 & cv ? & & & 23.48 & 23.44 & 26.01 & uncertain photometry + u28 & 17:40:38.904 @xmath9753:39:51.09 & 0.79 & & 1 & agn & ? & & 22.72 & 24.73 & 25.44 & + u31 & 17:40:34.202 @xmath9753:41:15.28 & 1.41 & cv11@xmath98 & 1 & cv & f & & 23.32 & 23.93 & 26.07 & + u41 & 17:40:45.008 @xmath9753:39:55.21 & 0.70 & & 1 & ms & c & c & 16.74 & 17.69 & 19.50 & only object in error circle + u42 & 17:40:43.059 @xmath9753:38:31.29 & 1.96 & v26 & 1 & ab & ? & & 15.52 & 17.47 & 18.20 & very red + u43 & 17:40:40.543 @xmath9753:40:22.79 & 0.26 & by - pc-4 & 1 & ab & c & c & 17.68 & 19.03 & 20.38 & + u60 & 17:40:47.807 @xmath9753:41:28.40 & 1.30 & cv9 & 1 & cv & c & c & 23.38 & 23.63 & 26.04 & + u61 & 17:40:45.223 @xmath9753:40:28.60 & 0.45 & cv12@xmath98 & 1 & cv ? & ? & & 22.08 & 23.19 & 24.19 & uncertain photometry + u62 & 17:40:30.422 @xmath9753:39:17.47 & 2.11 & & 1 & ab ? & & & 20.90 & 23.33 & 23.17 & uncertain photometry + u63 & 17:40:31.663 @xmath9753:38:46.36 & 2.31 & & 1 & ab ? & & & 18.15 & 20.39 & 20.72 & very red , moderate excess + u65 & 17:40:37.558 @xmath9753:39:17.85 & 1.36 & & 1 & ab ? & ? & & 18.72 & 21.27 & 21.11 & very red , large excess + u66 & 17:40:38.918 @xmath9753:38:49.80 & 1.71 & & 1 & ab ? & f & & 21.93 & 24.56 & 24.29 & red , large excess + u67 & 17:40:40.065 @xmath9753:40:16.59 & 0.37 & & 1 & ab & c & c & 19.34 & 21.17 & 22.00 & + u68 & 17:40:40.730 @xmath9753:38:32.63 & 1.95 & & 1 ? & ? & & & & & & marginal detection in @xmath0 only + u69 & 17:40:40.867 @xmath9753:40:17.17 & 0.27 & by - pc-5 & 1 & ab & c & c & 17.84 & 19.03 & 20.52 & + u70 & 17:40:41.693 @xmath9753:40:33.33 & 0.11 & v20 & 1 & ab , bs & c & c & 15.50 & 16.21 & 18.30 & + u73 & 17:40:42.681 @xmath9753:39:28.73 & 1.00 & & 1 & ab & c & c & 18.19 & 19.47 & 20.88 & + u75 & 17:40:43.666 @xmath9753:40:30.61 & 0.22 & & 1 & ab & c & & 19.71 & 21.78 & 22.32 & + u76 & 17:40:43.818 @xmath9753:41:16.32 & 0.83 & v17 & 1 & ab & c & c & 15.81 & 16.81 & 18.57 & + u77 & 17:40:44.125 @xmath9753:42:11.48 & 1.74 & v36 & 1 & ab ? & & & & & 24.21 & outside of @xmath0 field + u79 & 17:40:46.409 @xmath9753:40:04.05 & 0.75 & & 1 & ab & ? & & 19.82 & 21.98 & 22.41 & + u80 & 17:40:46.455 @xmath9753:41:56.50 & 1.60 & cv14@xmath98 & 1 & cv ? & & & 25.40 & 25.84 & 27.91 & weak detection + u81 & 17:40:46.481 @xmath9753:41:15.44 & 1.01 & v14 & 1 & ab & c & c & 18.62 & 20.27 & 21.29 & + u82 & 17:40:48.537 @xmath9753:39:39.53 & 1.25 & by - wf2 - 1 & 1 & ab & c & & 18.72 & 20.29 & 21.39 & + u83 & 17:40:49.615 @xmath9753:40:43.02 & 1.13 & cv15@xmath98 & 1 & cv ? & ? & ? & 26.22 & 27.18 & 28.56 & very faint + u84 & 17:40:54.807 @xmath9753:40:19.79 & 1.88 & & 1 & ab ? & ? & ? & & 17.27 & 18.23 & very red , outside of @xmath0 field + u86 & 17:40:37.473 @xmath9753:41:47.24 & 1.48 & & 1 & ab ? & ? & & 21.16 & 24.06 & 23.39 & very red , large excess + u87 & 17:40:42.877 @xmath9753:40:26.45 & 0.11 & by - pc-3 & 1 & ab & c & c & 17.46 & 18.69 & 20.14 & + u88 & 17:40:42.863 @xmath9753:40:23.43 & 0.13 & by - pc-1 , v19 & 1 & ab & c & c & 16.73 & 17.77 & 19.43 & + u89 & 17:40:43.613 @xmath9753:40:24.60 & 0.23 & & 1 & ab & c & c & 19.76 & 21.82 & 22.38 & + u90 & 17:40:41.779 @xmath9753:40:14.42 & 0.24 & by - pc-6 & 1 & ab & c & c & 18.45 & 19.88 & 21.11 & + u91 & 17:40:42.430 @xmath9753:40:41.65 & 0.22 & & 1 & ms & c & & 18.94 & 20.44 & 21.63 & only object in error circle + u92 & 17:40:43.916 @xmath9753:40:35.39 & 0.28 & by - wf4 - 2 , v7 & 1 & ab ? & c & c & 16.72 & 18.26 & 19.53 & very red , slight deficit + u93 & 17:40:42.393 @xmath9753:40:46.62 & 0.30 & & 1 & ? & & & & & 24.95 & uncertain photometry + u94 & 17:40:42.868 @xmath9753:40:49.07 & 0.36 & & 1 & ab & c & c & 18.88 & 20.58 & 21.53 & + u95 & 17:40:40.320 @xmath9753:40:44.58 & 0.38 & & 1 & ab ? & c & & 20.81 & 23.02 & 23.41 & ms color , excess + u96 & 17:40:39.097 @xmath9753:40:23.09 & 0.47 & v24 & 1 & ab & c & & 18.40 & 20.15 & 21.42 & + u97 & 17:40:43.918 @xmath9753:40:05.90 & 0.46 & & 1 & ? & c & & 20.90 & 22.88 & 23.56 & blue , normal + u98 & 17:40:40.994 @xmath9753:40:58.40 & 0.53 & & 1 & ab & c & & 19.18 & 21.09 & 21.82 & + u99 & 17:40:46.431 @xmath9753:40:30.40 & 0.63 & & 1 & ab & c & & 18.49 & 19.96 & 21.16 & + u100 & 17:40:38.201 @xmath9753:40:46.55 & 0.66 & & 1 & ab & c & & 20.81 & 23.14 & 23.40 & + u101 & 17:40:45.399 @xmath9753:41:01.40 & 0.73 & & 1 & ab ? & c & c & 18.48 & 19.94 & 21.19 & red , normal + u102 & 17:40:38.845 @xmath9753:39:43.12 & 0.90 & & 1 & ab & c & & 21.27 & 23.80 & 23.87 & + u103 & 17:40:35.698 @xmath9753:40:12.56 & 1.00 & & 1 & ab & c & & 19.26 & 21.05 & 21.93 & + u104 & 17:40:43.124 @xmath9753:39:29.04 & 1.01 & & 0 & & & & & & & empty error circle + u105 & 17:40:36.521 @xmath9753:41:07.85 & 1.06 & & 1 & ab & ? & & 19.78 & 21.84 & 22.37 & + u106 & 17:40:43.737 @xmath9753:39:17.52 & 1.21 & & 0 & & & & & & & empty error circle + u107 & 17:40:34.115 @xmath9753:40:17.01 & 1.21 & & 1 & ab & c & & 19.61 & 21.63 & 22.23 & + u108 & 17:40:52.099 @xmath9753:39:48.25 & 1.62 & & 1 & glx ? & & & & & & extended object + u109 & 17:40:52.728 @xmath9753:40:52.88 & 1.61 & & 1 & ab ? & c & c & 19.77 & 21.83 & 22.44 & red , normal + u110 & 17:40:33.455 @xmath9753:39:16.83 & 1.76 & & 1 & ab & c & & 19.92 & 22.04 & 22.53 & + u111 & 17:40:29.845 @xmath9753:40:26.99 & 1.82 & & 0 & & & & & & & empty error circle + u112 & 17:40:50.374 @xmath9753:39:06.00 & 1.84 & & 1 & ms & f & & 19.60 & 21.25 & 22.29 & only object in error circle + u113 & 17:40:42.764 @xmath9753:40:20.76 & 0.16 & & 0 & & & & & & & empty error circle + u114 & 17:40:43.469 @xmath9753:40:34.34 & 0.21 & & 1 & ? & c & & 19.49 & & 22.10 & uncertain photometry + u116 & 17:40:42.236 @xmath9753:40:19.97 & 0.18 & & 1 & ab & c & c & 18.18 & 19.49 & 20.88 & + u117 & 17:40:42.153 @xmath9753:40:25.56 & 0.14 & & 1 & ab & c & c & 16.82 & 17.84 & 19.55 & + u118 & 17:40:41.576 @xmath9753:40:15.88 & 0.05 & & 1 & msto & c & c & 15.60 & 16.60 & 18.34 & only object in error circle + u119 & 17:40:41.261 @xmath9753:40:19.35 & 0.22 & & 0 & & & & & & & empty error circle + u120 & 17:40:46.517 @xmath9753:40:15.64 & 0.21 & & 1 & ab ? & ? & ? & 21.09 & 23.70 & 23.39 & moderately red , large ha excess + u121 & 17:40:33.631 @xmath9753:39:34.96 & 0.68 & & 0 & & & & & & & empty error circle + u122 & 17:40:47.903 @xmath9753:39:24.83 & 1.55 & & 0 & & & & & & & empty error circle + u123 & 17:40:49.621 @xmath9753:38:45.93 & 2.04 & & 0 & & & & & & & empty error circle +

we have searched for optical identifications for 79 _ chandra _ x - ray sources that lie within the half - mass radius of the nearby , core - collapsed globular cluster ngc 6397 , using deep hubble space telescope advanced camera for surveys wide field channel imaging in , @xmath0 , and @xmath1 . photometry of these images allows us to classify candidate counterparts based on color - magnitude diagram location . in addition to recovering nine previously detected cataclysmic variables ( cvs ) , we have identified six additional faint cv candidates , a total of 42 active binaries ( abs ) , two millisecond pulsars ( msps ) , one candidate active galactic nucleus , and one candidate interacting galaxy pair . of the 79 sources , 69 have a plausible optical counterpart . the 15 likely and possible cvs in ngc 6397 mostly fall into two groups : a brighter group of six for which the optical emission is dominated by contributions from the secondary and accretion disk , and a fainter group of seven for which the white dwarf dominates the optical emission . there are two possible transitional objects that lie between these groups . the faintest cvs likely lie near the minimum of the cv period distribution , where an accumulation is expected . the spatial distribution of the brighter cvs is much more centrally concentrated than those of the fainter cvs and the active binaries . this may represent the result of an evolutionary process in which cvs are produced by dynamical interactions , such as exchange reactions , near the cluster center and are scattered to larger orbital radii , over their lifetimes , as they age and become fainter .

it was shown in ref . that the complex ernst potentials@xcite on the symmetry axis of an asymptotically flat , stationary and axisymmetric electro - vacuum spacetime surrounding a connected killing horizon ( including the case of a degenerate horizon ) can be constructed uniquely by means of the inverse scattering method@xcite . this provides a new proof of the kerr - newman black hole uniqueness since the full solution can straightforwardly be derived from the axis data , cf . here we present a particular method for obtaining the ernst potentials off the axis , see sec . the stationary and axisymmetric einstein - maxwell vacuum equations are equivalent to the ernst equations@xcite @xmath0 @xmath1 where @xmath2 and @xmath3 are weyl coordinates and a bar denotes complex conjugation . these equations are the integrability condition of a related linear problem ( lp)@xcite : @xmath4{\bf y}\ , , \ ] ] @xmath5{\bf y}\ ] ] @xmath6 @xmath7 @xmath8 where @xmath9 is the complex `` spectral parameter '' . on the upper part of the symmetry axis ( @xmath10 , @xmath11 ) , denoted by @xmath12 , a suitably normalized solution to the lp for the black hole problem ( horizon at @xmath10 , @xmath13 ) reads@xcite @xmath14 where we have assumed @xmath15 for @xmath16 . the ernst potentials @xmath17 and @xmath18 on @xmath12 are given by @xmath19 together with the parameter relation @xmath20 where @xmath21 , @xmath22 and @xmath23 denote mass , angular momentum and charge . the functions @xmath24 , @xmath25 , @xmath26 and @xmath27 are given by @xmath28 @xmath29 @xmath30 using the relation@xcite @xmath31 together with @xmath32 we are led from to the following ansatz for @xmath33 : @xmath34 with @xmath35 @xmath36 @xmath37 @xmath38 @xmath39 a careful discussion of the regularity conditions at @xmath40 leads to a set of linear algebraic equations which uniquely determine the unknowns @xmath41 , @xmath42 , @xmath43 , @xmath44 , @xmath45 and @xmath46 . according to , @xmath47 is given by @xmath48 . the resulting ernst potentials are the well - known expressions @xmath49 with boyer - lindquist coordinates @xmath50 and @xmath51 related to our weyl coordinates @xmath2 and @xmath3 by @xmath52

the kerr - newman black hole solution can be constructed straightforwardly as the unique solution to the boundary value problem of the einstein - maxwell equations corresponding to an asymptotically flat , stationary and axisymmetric electro - vacuum spacetime surrounding a connected killing horizon .

calculating qcd amplitudes by means of feynman diagrams can be an extremely challenging task . the gauge - dependence of vertices and unphysical degrees of freedom often makes intermediate steps immensely complicated . however , in 1998 chalmers and siegel showed that the complexity of feynman diagram calculations in yang - mills theory could be greatly reduced if a so - called space - cone gauge was used @xcite ( see also @xcite for similar simplifications in other theories ) . by now several alternative approaches are also available , such as the britto - cachazo - feng - witten ( bcfw ) recursion relation @xcite . at tree - level the above mentioned space - cone construction is closely related to these relations @xcite . the main goal with this short paper is to write down all feynman rules for qcd when working in the space - cone gauge . to our knowledge not all of these have been explicitly presented in the literature , and it is therefore our hope that this paper will allow for easy and straightforward application whenever such rules are needed . the paper is structured as follows ; in section [ sec : yml ] we review the yang - mills lagrangian in space - cone gauge and the elimination of unphysical degrees of freedom . section [ sec : spf ] introduces some notation and the spinor helicity formalism . in section [ sec : ymf ] we give the feynman rules following from section [ sec : yml ] . in section [ sec : bcfw ] we make some comments on the connection between bcfw relations and the space - cone gauge . in section [ sec : af ] we add quarks to the lagrangian and show that effective four - point vertices involving quark - antiquark pairs will appear . in section [ sec : qf ] we give the feynman rules for quarks and finally in section [ sec : con ] we have our conclusions . we start from the standard lagrangian of yang - mills theory @xmath0 = -\frac{1}{4}f^a_{\mu\nu}f^{\mu\nu\ , a } , \label{ym_lag}\ ] ] where @xmath1 = if_{abc}t^c , & \nonumber \\ & \mathrm{tr}[t^at^b ] = \frac{1}{2}\delta^{ab},&\end{aligned}\ ] ] and @xmath2 the contraction gives @xmath3 here @xmath4 is just the usual vector field with inner product given by @xmath5 we now introduce the lightcone components @xmath6 and in terms of these the inner product is @xmath7 our first goal is to express eq . in terms of the lightcone components , however , we use the gauge freedom to set @xmath8 , and hence discard all terms containing an @xmath9 . note that in many of the intermediate calculations we have used the fact that a term symmetric in two color indices will vanish when contracted with the anti - symmetric colorfactor . written in terms of the lightcone components the quadratic part of eq . becomes @xmath10 \nonumber \\ & & -2\left [ \partial^-a^{+\,a}\partial^-a^{+\,a } + \partial^+a^{-\,a}\partial^+a^{-\,a } + \partial\bar{a}^a\partial\bar{a}^a \right],\end{aligned}\ ] ] the three - point interaction @xmath11 and the four - point interaction @xmath12 collecting these expressions the lagrangian takes the form @xmath13 \nonumber \\ & -g f_{abc } \left [ ( \partial^-a^{+\,a})a^{+\,b}a^{-\,c } + ( \partial^+a^{-\,a})a^{-\,b}a^{+\,c } \right . \nonumber \\ & \phantom{aaaaa}\left . -(\partial a^{+\,a})\bar{a}^ba^{-\,c } - ( \partial a^{-\,a})\bar{a}^ba^{+\,c}\right ] \nonumber \\ & -\frac{1}{2}g^2f_{abc}f_{ab'c'}a^{+\,b}a^{-\,c}a^{-\,b'}a^{+\,c'}\ , . \label{lbefore}\end{aligned}\ ] ] following @xcite we then use the equation of motion for @xmath14 to eliminate this `` auxiliary '' component from the lagrangian , that is , we use @xmath15 and get the following expression for @xmath16 @xmath17.\end{aligned}\ ] ] plugging this back into eq . , and doing a bit of rewriting , we obtain @xmath18 this is the pure yang - mills lagrangian written in terms of two scalar fields @xmath19 and @xmath20 , consistent with massless vector fields only having two physical degrees of freedom . we choose to use the pauli matrices with the following normalization @xmath21 such that a contraction between these and a four - vector is @xmath22 and @xmath23 note that @xmath24 , so if @xmath25 this matrix only has one non - vanishing eigenvalue and can be decomposed into a bispinor product @xmath26 we will from now on use the following braket notation @xmath27 notice that our convention differs from @xcite . as always when one wants to do amplitude calculations by feynman rules there are some lorentz frames in which the calculations are easier to perform . to set up this we introduce two ( for now ) arbitrary massless _ reference _ momenta , one denoted with a @xmath28 and one denoted with @xmath29 . we then choose to work in the lorentz frame where our two reference momenta have the following simple expressions @xmath30 that is the frame where they both move along the @xmath31 axis , but in opposite direction . note that @xmath32 , and that we have called the spinor of the @xmath28 momentum for @xmath33 and vice versa . our choice of labelling will soon become apparent . since any four - vector , contracted with the pauli matrices , is written in the form of eq . , using the following normalized matrices as basis @xmath34 four - vectors take the form @xmath35 where the coefficients are just the lightcone components . since @xmath36 the lightcone components are @xmath37 we also choose to use our reference momenta in the expression for the polarization vectors @xmath38 @xmath39 that is , we have used our reference momenta to fix the gauge - freedom one has in polarization vectors . from eq . it is evident that we are only concerned with the @xmath40 components , and from eq . we see that @xmath41 hence , with this setup we have that only the positive helicity gluons have the @xmath42 lightcone component and only the negative helicity gluons the @xmath43 component . for this reason the @xmath44 labels in eq . is now actually denoting the helicity and not just the specific lightcone component . in this section we present the color - ordered feynman rules one obtains from eq . , that is the rules one could use in calculating , for instance , the partial tree amplitudes @xmath45 in @xmath46 a_n(1,2,\ldots , n)\,,\end{aligned}\ ] ] where the sum is over all non - cyclic permutations of external legs . before we do so , remember that in the last section our massless reference momenta was just arbitrarily chosen , however , if we make the choice that @xmath47 is one of the _ external _ momenta of a @xmath42 helicity gluon and @xmath48 one of the _ external _ momenta of a @xmath43 helicity gluon , the rules and explicit calculations simplify greatly . the external _ non_-reference legs will just contribute with the plus or minus lightcone component of the polarization vector , depending on the helicity ( note that we _ always _ take external momenta to be outgoing ) , _ i.e. _ @xmath49 or @xmath50 however , for the reference legs the @xmath51 vanish because of the @xmath52 and @xmath53 in the numerator . these factors can only be countered in the three - point vertex and only if they sit on the @xmath54 term , _ i.e. _ @xmath55 the product of these two contributions is @xmath56 , and since every diagram will always contain this product ( assuming one takes both reference momenta to correspond to external legs ) , we just write the external lines for reference legs as @xmath57 the term representing the propagator is the usual boson propagator ( notice that we for simplicity discard all factors of @xmath58 in the following rules ) @xmath59 the three - point vertex splits into the case where one of the lines is a reference leg and the case in which non of the lines are . with a reference leg present we have used up the @xmath54 term for a cancellation like in eq . and are only left with the contribution from the opposite helicity leg , through @xmath60 , _ i.e. _ + @xmath61 or + @xmath62 + + where the minus sign in the second diagram comes from the antisymmetry of the colorfactor @xmath63 . there can only be _ one _ reference leg on a three - point vertex since there is only one term in these that can counter the vanishing of that leg . in the second case of no reference leg , the three - point vertex is + @xmath64 + + again the minus sign is a consequence of the anti - symmetric colorfactor . we can never have a reference leg on a four - point vertex since these do not contribute with a similar `` @xmath65 '' counter - term . however , there are still two different kinds of four - point vertices because of the two possible helicity configurations @xmath66 and @xmath67 ( cyclically speaking ) . the first case is + @xmath68 + + and the second case + + @xmath69 + these rules reduce the number of diagrams contributing to a specific color - ordered amplitude significantly . however , the reduction relies heavily on the choice of having external momenta as reference momenta such that diagrams with both reference legs on the same three - point vertex and diagrams with a reference leg on a four - point vertex all vanish . the rules are of course perfectly allowed without this choice , but then no such constraints exist and the simplified diagrams , and should be discarded . before turning to the inclusion of quarks let us make some comments on the connection between working in space - cone gauge and the bcfw recursion relation . one of the interesting features of the space - cone gauge is , that a diagrammatic proof of the bcfw recursion relation can be obtained from the above feynman rules . the main observation is that no vertex depends on @xmath70 . therefore , a shift in the @xmath71 direction of a reference leg leaves no imprint on the vertices , but only on internal propagators . using a propagator relation between shifted and unshifted propagators , it is possible to relate amplitudes calculated in space - cone gauge to the bcfw result . for the four- and five - point case this calculation was done in @xcite where the reader may also find details regarding the propagator identity . however , these two examples are in a way special , since no four - point vertex enters the calculations . the first case where the four - point vertex is manifest for all choices of reference momenta is the @xmath72 six - point amplitude . in a very condensed notation it is given by @xmath73 where @xmath28 and @xmath29 denotes the positive and negative reference gluons respectively . when we use the propagator identity introduced by vaman and yao , we obtain a factorization which has the structure of a three - point amplitude times a five - point amplitude . here the four - point vertex enters explicitly in the five - point amplitude , which seems to be in conflict with @xcite , where it was shown explicitly that the five - point amplitude is independent of the four - point vertex . @xmath74 @xmath75 to explain this , we notice , that the result in @xcite exploits the fact that in the space - cone gauge , the number of diagrams and the types of vertices that enters a specific calculation , is highly dependent on the chosen reference lines . when two of the external gluons are chosen as references , the four - point vertex in the five - point amplitude is absent . however , if we only choose one of the external momenta as reference , the four - point vertex will still be present in the calculation . the structure of the five - point amplitude , with only one external gluon chosen as reference line , is shown below . @xmath76 comparing this with and we notice , that the diagrams with four - point vertices exactly matches those of . the diagrams with only three - point vertices can just as easily be found when using the propagator identity in the rest of the diagrams . we therefore conclude that the amplitudes combine in exactly the right way as to reproduce the bcfw result , and that we should expect the procedure to generalize to any number of external legs . lastly , we would like to make a comment regarding the propagator identity that is needed in order to conclude the bcfw recursion relation in @xcite . it is shown that the identity is equivalent to @xmath77 over a contour enclosing all poles . in the original proof of the bcfw recursion relation @xcite , two main properties of tree - level amplitudes was needed . first that the poles of the shifted amplitude @xmath78 all come from propagators going on - shell . second , that @xmath79 as @xmath80 . these statements are proved in @xcite , but the arguments become almost trivial when put into the light of the space - cone formalism . the first statement is immediately clear since the shifts can be chosen such that the vertices of the amplitude are unaffected . hence , only internal propagators change . then the second statement follows from since the integrand obviously converges fast enough to zero . we will now add quarks to our lagrangian . this amounts to adding @xmath81 to eq . , where @xmath82 is just the lagrangian for the free dirac field , and the interaction can be written out in terms of the lightcone components as @xmath83\,.\end{aligned}\ ] ] note that we have also written the @xmath84-matrices in the lightcone notation now , _ i.e. _ @xmath85 , _ etc . _ the expression for @xmath16 , following from eq . , should then be replaced by @xmath17 \nonumber \\ & -g\frac{1}{\partial^2 } \left ( \bar{\psi } \gamma t^a \psi \right)\,,\end{aligned}\ ] ] and after substituting this back into the original lagrangian we obtain @xmath86 where @xmath87 \nonumber \\ & + g \left [ a^{-a}\bar{\psi } \gamma^+ t^a \psi - \left(\frac{\partial^+}{\partial } a^{-a } \right ) \bar{\psi}\gamma t^a \psi \right ] \nonumber \\ & + g^2 f_{abc } \left [ \frac{1}{\partial } \left ( ( \partial a^{+a } ) a^{-c } \right ) \frac{1}{\partial } \left ( \bar{\psi}\gamma t^b\psi\right ) \right . \nonumber \\ & \phantom{aaaaaa}\left . - \frac{1}{\partial } \left ( ( \partial a^{-c } ) a^{+a } \right ) \frac{1}{\partial } \left ( \bar{\psi}\gamma t^b\psi\right)\right ] \nonumber \\ & -\frac{g^2}{2 } \frac{1}{\partial } \left ( \bar{\psi}\gamma t^a \psi \right ) \frac{1}{\partial}\left ( \bar{\psi}\gamma t^a \psi \right)\ , . \label{l_i}\end{aligned}\ ] ] the external states for quarks are given by @xmath88 and @xmath89 where @xmath90 , and @xmath91 and @xmath92 are positive- and negativ - energy solutions , respectively , of the dirac equation , with @xmath93 labelling spin up or down . the propagator is just the usual dirac propagator @xmath94 from eq . we can read off the color - ordered feynman rules involving vertices with quarks . the three - point vertices involving one gluon and a quark - antiquark pair is + + @xmath95 + and + @xmath96 + or in case of a reference leg + + @xmath97 + and + @xmath98 due to the elimination of the @xmath14 component from the lagrangian , we have instead introduced four - point vertices involving quarks . the first one is an effective gluon - gluon - quark - antiquark interaction + @xmath99 + and + @xmath100 like in the pure gluon case there can not be reference legs on these four - point vertices either . the second one is a double pair quark - antiquark interaction + @xmath101 + and + @xmath102 + notice that the @xmath84-matrices in eq . and are _ not _ multiplied together . in a real amplitude calculation they will appear between corresponding external spinors , for example like + @xmath103 this covers all the feynman rules one obtains for qcd in the space - cone gauge . in this paper we have explicitly written down all feynman rules for qcd in space - cone gauge when unphysical degrees of freedom in the gluonic sector have been removed . combined with a clever choice of reference frame this reduces the amount of feynman diagrams needed for gluon amplitude calculations considerably . we then made some comments about the close connection between bcfw recursion relations and the space - cone gauge , especially concerning the role played by the four - point vertex . we have also seen that in the presence of quarks the former manipulations of the lagrangian lead to the introduction of effective four - point interaction terms involving quark - antiquark pairs . g. chalmers and w. siegel , phys . d * 59 * ( 1999 ) 045013 [ hep - ph/9801220 ] . g. chalmers and w. siegel , phys . d * 59 * ( 1999 ) 045012 [ hep - ph/9708251 ] . g. chalmers and w. siegel , phys . d * 63 * ( 2001 ) 125027 [ hep - th/0101025 ] . r. britto , f. cachazo and b. feng , nucl . phys . b * 715 * , 499 ( 2005 ) [ hep - th/0412308 ] . r. britto , f. cachazo , b. feng and e. witten , phys . * 94 * , 181602 ( 2005 ) [ hep - th/0501052 ] . d. vaman and y. -p . yao , jhep * 0604 * ( 2006 ) 030 [ hep - th/0512031 ] .

we present the lagrangian and feynman rules for qcd written in space - cone gauge and after eliminating unphysical degrees of freedom from the gluonic sector . the main goal is to clarify and allow for straightforward application of these feynman rules . we comment on the connection between bcfw recursion relations and space - cone gauge .

this paper surveys the work in recent years on the classification problem for partially hyperbolic dynamical systems . one motivation for research into partial hyperbolicity is the study of invariant objects in dynamical systems which persist under perturbation . to this end , we start the survey by defining isolated sets and investigating their properties under changes to the dynamics . consider a diffeomorphism @xmath0 on a riemannian manifold . a compact set @xmath1 is _ invariant _ if @xmath2 . it is _ isolated _ if there is a neighborhood @xmath3 such that @xmath4 note that any isolated set is automatically invariant . let @xmath5 denote the space of diffeomorphisms @xmath0 equipped with the @xmath6 topology @xcite . for an isolated set @xmath7 and @xmath8 close to @xmath9 , the _ continuation _ of @xmath10 is defined by @xmath11 . in general , the continuation @xmath12 could be very different in structure from the original @xmath10 . the simplest example is where @xmath10 is a singleton set @xmath13 . one can verify that @xmath12 is still a singleton set for all nearby @xmath14 exactly when the derivative @xmath15 is hyperbolic ( i.e. , with no eigenvalues of modulus one ) . consider now the case where the invariant set @xmath1 is a submanifold . let @xmath16 denote the tangent bundle @xmath17 restricted to @xmath1 . a _ subbundle _ @xmath18 is a continuous choice of subspace @xmath19 for each @xmath20 . the subbundle is _ invariant _ if @xmath21 where @xmath22 is the derivative of @xmath9 . for example , if @xmath23 is an embedded submanifold , then @xmath24 , the set of vectors tangent to @xmath1 , is an invariant subbundle of @xmath25 . an invariant submanifold @xmath1 is _ normally hyperbolic _ if there is a splitting of @xmath25 into invariant subbundles @xmath26 and @xmath27 such that at each @xmath20 the inequalities @xmath28 hold for all unit vectors @xmath29 , @xmath30 and @xmath31 . in words , this means that the expansion and contraction normal to the submanifold dominates any expansion or contraction tangent to the submanifold . an isolated @xmath32-submanifold @xmath10 is called _ persistent _ if @xmath12 is also a @xmath32-submanifold close to the original @xmath10 for all @xmath14 in a @xmath32 neighborhood of @xmath9 . [ thm : nhper ] normally hyperbolic submanifolds are persistent . this result , in one form or another , has been proved many times over , with the techniques of the proof going back at least to a 1901 paper of hadamard on stable manifolds . see ( * ? ? ? * section 1 ) for references . in his 1973 thesis , r. ma @xcite proved the converse . [ ma ] [ thm : pernh ] persistent submanifolds are normally hyperbolic . analogous results also hold for continuous - time dynamical systems @xcite . in their 1977 book , hirsch , pugh , and shub developed extensions to [ thm : nhper ] in great detail @xcite . one case considered is when @xmath1 supports a lamination . that is , @xmath1 decomposes as a disjoint union of submanifolds , called leaves , such that if @xmath33 denotes the leaf through @xmath20 , then the tangent space @xmath34 depends continuously on @xmath35 . this lamination is invariant if both @xmath2 and @xmath36 for all @xmath20 . we use @xmath37 to denote the lamination and say that @xmath10 is the support of @xmath37 . normal hyperbolicity is then defined as above with @xmath38 in place of @xmath39 . hirsch , pugh , and shub proved a limited form of persistence for normally hyperbolic laminations ( * ? ? ? * theorem 6.8 ) , but a priori the lamination structure may be destroyed under perturbation . for instance , if @xmath33 and @xmath40 are distinct leaves in @xmath37 , then the corresponding submanifolds in @xmath12 may a priori intersect without coinciding . if we add one additional assumption called `` plaque expansiveness , '' then these thorny issues disappear . as it is technical , we omit the definition of plaque expansiveness . to state the theorem , however , we define one more notion . suppose @xmath41 are diffeomorphisms with respective invariant laminations @xmath37 and @xmath42 . a _ leaf conjugacy _ between @xmath43 and @xmath44 is a homeomorphism @xmath45 such that if @xmath33 is a leaf of @xmath37 , then @xmath46 is a leaf of @xmath42 and @xmath47 . that is , conjugacy holds on the level of leaves . [ thm : peconj ] suppose @xmath0 is a diffeomorphism and @xmath37 is an @xmath9-invariant , normally hyperbolic , plaque expansive lamination . then , there is a @xmath32-neighborhood @xmath48 of @xmath9 such that for every @xmath49 there is a @xmath14-invariant , normally hyperbolic , plaque expansive lamination @xmath42 and @xmath44 is leaf conjugate to @xmath43 . plaque expansive holds in every known example . is every normally hyperbolic lamination also plaque expansive ? what are all possible normally hyperbolic laminations ? can they be classified up to some notion of equivalence ? in light of [ thm : peconj ] the most natural notion of equivalence would be leaf conjugacy . this problem may be viewed as the major motivation of the research detailed in this survey paper . as the problem in its full generality is way too difficult at present , we look at it in specific settings . normal hyperbolicity builds on the notion of uniform hyperbolicity developed by anosov , smale , and others in the 1960s . consider a diffeomorphism @xmath0 . a compact invariant subset @xmath1 is _ hyperbolic _ if there is an invariant splitting @xmath50 and @xmath27 such that @xmath51 for all unit vectors @xmath52 and @xmath53 . a hyperbolic set may be viewed as a special case of a normally hyperbolic lamination , where the `` leaf '' through a point @xmath35 is just the singleton set @xmath54 . the leaf conjugacy in [ thm : peconj ] is then a true conjugacy between the homeomorphisms @xmath55 and @xmath56 . further , `` plaque expansiveness '' is just `` expansiveness '' here , and is always known to hold for hyperbolic sets . when the manifold @xmath57 is a hyperbolic set , the diffeomorphism @xmath0 is called _ anosov_. several important classification results have been established for anosov diffeomorphisms and we discuss these in [ sec : classanosov ] . shortly after the development of normal hyperbolicity , brin and pesin developed a related notion of partial hyperbolicity @xcite . there are several `` flavours '' of partial hyperbolicity . as this is a survey on the subject , we give a list of the various definitions . consider again a diffeomorphism @xmath0 and assume there is a compact invariant set @xmath1 with invariant splitting @xmath58 such that at each @xmath20 the inequalities @xmath59 hold for all unit vectors @xmath29 , @xmath60 and @xmath31 . we assume @xmath61 is non - zero , as otherwise the set would be uniformly hyperbolic . if exactly one of the bundles @xmath62 and @xmath63 is non - zero , then @xmath1 is a _ weakly partially hyperbolic set_. if @xmath62 , @xmath61 , and @xmath63 are all non - zero , then @xmath1 is a _ strongly partially hyperbolic set_. the bundles @xmath62 , @xmath61 , and @xmath63 are called the `` stable '' , `` center '' , and `` unstable '' bundles respectively . note that any strongly partially hyperbolic set may also be viewed as a weakly partially hyperbolic by , say , grouping the stable and center bundles together into a single center bundle . the definition of partial hyperbolicity above is pointwise ; the inequalities hold independently at each @xmath20 . a partially hyperbolic set is _ absolutely _ partially hyperbolic if the inequalities hold for any points @xmath64 and unit vectors @xmath29 , @xmath65 , and @xmath66 . recently , rodriguez hertz , rodriguez hertz , and ures constructed interesting examples of pointwise partially hyperbolic systems that are impossible in the absolute setting @xcite . throughout this paper , `` partially hyperbolic '' means strongly pointwise partially hyperbolic unless otherwise noted . we also primarily consider dynamical systems where the partially hyperbolic set is the entire manifold @xmath67 and say that the diffeomorphism @xmath0 is partially hyperbolic . in this setting , one can ask if there is an invariant foliation @xmath68 tangent to the center direction @xmath61 . such a foliation would necessary be normally hyperbolic . we call a partially hyperbolic diffeomorphism _ dynamically coherent _ if there are invariant foliations @xmath69 , @xmath70 , and @xmath68 tangent respectively to @xmath71 , @xmath72 , and @xmath61 . which partially hyperbolic systems are dynamically coherent ? can the dynamically coherent systems be classified up to leaf conjugacy ? can the non - dynamically coherent systems be classified up to some other notion of equivalence ? partially hyperbolic dynamics arise naturally when studying robustly transitive systems . an invariant set is _ transitive _ if it contains a dense orbit . that is , if there is @xmath20 such that @xmath73 an isolated set is @xmath6-_robustly transitive _ if the continuation @xmath12 is transitive for all @xmath14 in a @xmath6-neighborhood of @xmath9 . transitive hyperbolic sets are robustly transitive and form a large class of examples . ma showed that in dimension 2 , the two notions are equivalent @xcite . suppose @xmath57 is a surface and @xmath0 a diffeomorphism . a transitive set @xmath1 is @xmath32-robustly transitive if and only if it is hyperbolic . higher dimensions are more complicated . in 1968 , shub constructed a robustly transitive example on @xmath74 which is partially hyperbolic but not hyperbolic ( see ( * ? ? ? * section 8) ) . in 1978 , ma constructed a similar example on @xmath75 @xcite . in 1995 , bonatti and daz developed the idea of a `` blender '' to construct many examples of invariant sets that are both partially hyperbolic and robustly transitive @xcite . in 1999 , daz , pujals , and ures proved the following theorem in the converse direction @xcite . for a diffeomorphism on a 3-dimensional manifold @xmath57 , any @xmath32-robustly transitive set must be weakly partially hyperbolic . what are the robustly transitive sets in dimension 3 ? in particular , if @xmath57 is a compact 3-manifold @xmath57 , then for which diffeomorphisms @xmath0 is the entire manifold @xmath57 a robustly transitive set ? can any 3-manifold @xmath57 support such a diffeomorphism @xmath9 ? robustly transitive sets in dimensions 4 and higher must also have some form of invariant splitting @xcite called `` volume partial hyperbolicity '' or just `` volume hyperbolicity . '' this notion is similar in form to partial hyperbolicity , but there are robustly transitive sets in these higher dimensions which are not strictly partially hyperbolic . see ( * ? ? ? * chapter 7 ) . interesting results also hold in the continuous time setting . for flows in dimension 2 , the only robustly transitive sets are equilibrium points and attracting and repelling orbits @xcite . for flows in dimension 3 , every @xmath32-robustly transitive set must be weakly partially hyperbolic . examples include the lorenz attractor , and any robustly transitive set with an equilibrium point can , in some sense , be viewed as a finite number of geometric lorenz attractors `` glued together . '' for further details and precise statements see @xcite or ( * ? ? ? * chapter 9 ) and references therein . finally , we mention stable ergodicity as a major motivation for studying partially hyperbolic dynamics . invariant measures often arise naturally in dynamics , for instance , when studying symplectic dynamical systems and specifically in the case of hamiltonian systems due to liouville s theorem . consider a diffeomorphism @xmath9 of a compact manifold @xmath57 and suppose there is an invariant probability measure @xmath76 on @xmath57 . this pair @xmath77 is _ ergodic _ if @xmath78 for almost every @xmath35 . in other words , @xmath76 captures the average behaviour of almost every orbit of @xmath9 . for this discussion , we assume @xmath76 is equivalent to lebesgue measure on @xmath57 . the pair @xmath77 is _ stably ergodic _ if @xmath79 is also ergodic for every @xmath76-preserving diffeomorphism @xmath14 close to @xmath9 . ( for reasons lost to history , the term `` stably ergodic '' has won out over `` robustly ergodic . '' ) in the 1960s , d. v. anosov demonstrated that volume - preserving anosov diffeomorphisms are stably ergodic @xcite . this is a large part of the reason that they are called `` anosov '' diffeomorphisms . for decades , anosov systems were the only known examples of stable ergodicity , though some systems were known to be stably ergodic among certain interesting classes of perturbations . a notable case is that of frame flows in negative curvature which represent a class of genuinely partially hyperbolic flows , which under certain conditions can be shown to be stably ergodic among frame flows @xcite . in 1995 , grayson , pugh , and shub gave an example of a partially hyperbolic diffeomorphism which is stably ergodic @xcite.this lead to a frenzy of research in the subsequent years showing that many volume - preserving partially hyperbolic systems are stably ergodic . see @xcite for surveys of the results relating partial hyperbolicity and stable ergodicity . see also @xcite which directly apply classification and leaf conjugacy results to the study of ergodicity . many surveys on partial hyperbolicity already exist @xcite . particularly , a recent survey treats partially hyperbolic diffeomorphisms in dimension 3 including a considerable section devoted to classification and integrability @xcite . it is therefore a big challenge to write a survey which complements these existing materials . we felt that there were still interesting things to be said , in particular regarding the classification problem . in this paper , many relevant aspects are omitted or only briefly discussed if they are covered in detail in other surveys . for instance , it is standard when surveying a mathematical subject to start with a list of known examples . instead , we refer the reader to the above surveys and to ( * ? ? ? * section 3 ) . a geometric structure such as a splitting of the tangent bundle into subbundles will likely impose some form of restriction both on the manifold and on the isotopy class of a diffeomorphism which leaves it invariant . for anosov diffeomorphisms in dimensions 4 or higher , however , the classification problem is yet completely open . for anosov flows , the same problem arises . even in dimension 3 , where many results are known , there are still many open questions . for example , infinitely many hyperbolic 3-manifolds are known to support anosov flows @xcite and infinitely many do not @xcite , but there is still an infinite class of hyperbolic manifolds where the question is wide open . classifying partially hyperbolic systems is , in principle , much more complicated and includes the classification of anosov diffeomorphisms and flows as subproblems . indeed , the understanding of which manifolds and isotopy classes admit partially hyperbolic systems is much less advanced than for anosov systems . the bulk of this survey concerns the classification of partially hyperbolic diffeomorphisms in closed 3-dimensional manifolds . higher dimensions are discussed only briefly at the end . the main point , suggested informally by pujals in a conference and later formalized by bonatti and wilkinson @xcite is that even if classification of anosov systems may be out of reach , it is reasonable to attempt to _ compare _ partially hyperbolic diffeomorphisms to anosov systems . many conjectures and perspectives were triggered by this approach and we refer to the recent survey @xcite . here we present a series of recent results and examples related to this approach . of course , the choice of results is biased by the interests of the authors with a special emphasis on presenting both our previous and ongoing joint work . there have been essentially two approaches to the classification problem . the first , initiated in @xcite , looked at the behavior of the center foliation in known examples and showed that this behavior characterized the examples . this has been quite successful and it has even lead to progress in higher dimensions @xcite and it is partially surveyed in [ sec : others ] . a different approach , initiated by the work of brin , burago and ivanov @xcite , instead makes assumptions on the underlying topology of the manifold and , in cases , also on the isotopy class of the diffeomorphism . in the rest of this section , we restrict to the 3-dimensional case . the study of 3-manifolds is quite advanced , and it makes sense to separate the study of partially hyperbolic diffeomorphisms based on the geometry of the manifold . first , we consider `` small '' manifolds where the volume of a ball in the universal cover is polynomial in the radius of the ball . equivalently , the fundamental group has polynomial growth ( see @xcite ) . in this setting , unstable curves grow exponentially fast under the dynamics and in some sense wrap around themselves in the compact manifold . this idea of curves in manifold `` wrapping around themselves '' under iteration may be made rigorous by considering the action in homology . this was formalized and proved in @xcite and extended in @xcite to show the following : [ thm : phpoly ] let @xmath9 be a partially hyperbolic diffeomorphism supported on a 3-manifold @xmath57 such that the fundamental group has polynomial growth . then , ( up to finite cover ) @xmath57 is a circle bundle over a torus , and the action @xmath80 on homology @xmath81 has eigenvalues @xmath82 with @xmath83 . give an outline of results featuring in the proof of [ thm : phpoly ] . circle bundles over the torus may also be viewed as quotients of nilpotent lie groups by a discrete lattice . that is , they are nilmanifolds . further , for every nilmanifold in dimension three there is an automorphism of the nilpotent lie group which quotients down to a partially hyperbolic diffeomorphism on the nilmanifold ( * ? ? ? * appendix a ) . using this , the authors were able to classify all partially hyperbolic diffeomorphisms on these manifolds @xcite . ( see also @xcite for previous results . ) [ thm : nil ] let @xmath9 be a partially hyperbolic diffeomorphism supported on a 3-manifold @xmath57 such that the fundamental group is nilpotent and suppose there is no compact surface tangent to @xmath71 or @xmath72 . then , ( up to finite cover ) @xmath9 is leaf conjugate to an algebraic map on @xmath57 . we next consider the case where the fundamental group has exponential growth . this comprises a much larger class of 3-manifolds . further , exponential growth is a necessary condition for a manifold to support an anosov flow @xcite . in this class of `` big '' manifolds , the panorama is less clear . in this setting , the simplest manifolds to study are those with solvable fundamental group . these are the manifolds produced by the suspension of an anosov diffeomorphism , and the authors showed that the following classification result holds @xcite . [ thm : sol ] let @xmath9 be a partially hyperbolic diffeomorphism supported on a 3-manifold @xmath57 such that the fundamental group is solvable and has exponential growth . further , suppose there is no compact surface tangent to @xmath71 or @xmath72 . then , an iterate of @xmath9 is leaf conjugate to the time - one map of a suspension anosov flow . the main goal of this survey is to outline ideas in the proofs of [ thm : nil , thm : sol ] and the results on which they rely . details classification results for foliations on 3-manifolds with solvable fundamental group . gives an overview of the techniques to establish leaf conjugacy which are common to all of the proofs . the proof of [ thm : nil ] is split into three special cases , handled in [ sec : skew , sec : da , sec : nil ] . finally , [ thm : sol ] is reviewed in [ sec : sol ] . after manifolds with solvable fundamental group , the next family of manifolds in terms of complexity are seifert ( fiber ) spaces . these generalize circle bundles over surfaces . ghys showed that , up to finite cover , any anosov flow on a circle bundle is orbit equivalent to a geodesic flow @xcite . barbot generalized this result to classify anosov flows on all seifert spaces @xcite . in recent work with m. shannon , the authors established the following @xcite . [ thm : phseifert ] let @xmath57 be a seifert space whose fundamental group has exponential growth . then , @xmath57 admits a transitive partially hyperbolic diffeomorphism if and only if it admits an anosov flow . [ cor : noproduct ] if @xmath84 is a surface of genus @xmath85 , then @xmath86 does not admit a transitive partially hyperbolic diffeomorphism . we present an idea of the proofs in [ sec : seifert ] . note that [ thm : phseifert ] does not provide a classification up to leaf conjugacy . at the moment of writing , this seems to be a difficult problem related to the recent discovery of new examples on seifert spaces which are not isotopic to the identity @xcite . we survey these examples in [ ss - examples ] along with other examples @xcite which must be taken into account if one hopes to establish a classification up to leaf conjugacy in full generality . other classification results are presented in section [ sec : others ] . as we present the main ideas behind these results and their proof , we also state several questions to help readers grasp the ongoing nature of this program . hopefully , it serves to interest readers and prompts them to join in this exciting quest . a key tool in classification results for both hyperbolic and partially hyperbolic systems is the franks semiconjugacy presented in this section . we also briefly describe the classification problem for anosov diffeomorphisms . [ franks ] [ thm : franks ] suppose @xmath87 is a linear hyperbolic toral automorphism and @xmath0 and @xmath88 are continuous maps such that @xmath89 commutes . then , there is a continuous map @xmath90 homotopic to @xmath91 such that @xmath92 commutes . if @xmath93 is non - zero , then @xmath94 is surjective . since ideas arising in the proof are fundamental to later results for partially hyperbolic systems , we give a sketch of the proof . throughout this survey , we present `` ideas of proofs '' in order to give the reader an intuition as to why a result might hold . these should not be regarded as rigorous proofs , and the reader is most welcome to skip over them . for simplicity , we only consider the case @xmath95 and where @xmath96 is the identity map . then , @xmath97 lifts to a map @xmath98 which is a finite distance from the lifted linear map @xmath99 . as @xmath100 is hyperbolic , there is a linear unstable foliation @xmath101 of @xmath102 where each leaf @xmath103 is a translate of the unstable eigenspace of @xmath100 . the set @xmath104 is a complete metric space when equipped with the metric @xmath105 as @xmath106 and @xmath100 are at finite distance , the function @xmath107 is well defined . if @xmath108 is the stable eigenvalue of @xmath100 , then @xmath109 for any two ( linear ) leaves @xmath110 and thus @xmath111 for any @xmath112 . as @xmath113 is a contraction , it has a unique fixed point @xmath114 . this function @xmath115 assigns to each point @xmath35 the unique leaf @xmath116 such that @xmath117 is bounded for all @xmath118 . note that if @xmath119 is a closed , non - empty subset of @xmath120 and @xmath121 , it follows that @xmath122 . from this one can show that @xmath123 is continuous and commutes with every deck transformation @xmath124 . an analogous function @xmath125 exists . define @xmath126 . then @xmath127 satisfies @xmath128 and quotients down to a semiconjugacy on @xmath129 . the above theorem may be used to find a semiconjugacy @xmath94 from a non - linear anosov systems @xmath9 to its linear part @xmath100 . to establish that @xmath94 is a true topological conjugacy between @xmath9 and @xmath100 , one must also show that @xmath94 is invertible . franks showed that under certain assumptions for @xmath9 , the above proof may be adapted to construct a semiconjugacy in the opposite direction . that is , to find @xmath130 where @xmath131 and then @xmath132 . he used this to give a classification of anosov diffeomorphisms on surfaces @xcite . he also gave specific conditions for classification that were later verified by s. newhouse @xcite and a. manning @xcite to establish the following results . [ franks newhouse ] if @xmath0 is an anosov diffeomorphism with @xmath133 , then @xmath57 is homeomorphic to a torus @xmath134 . [ franks manning ] [ thm : franksmanning ] if @xmath135 is an anosov diffeomorphism , then @xmath9 is topologically conjugate to a hyperbolic toral automorphism @xmath87 . the franks manning result also holds for anosov systems supported on nilmanifolds and infranilmanifolds showing they are conjugate to simple algebraic examples . the fundamental question is if there are any other possibilities . is every anosov diffeomorphism topologically conjugate to an algebraic map on a infranilmanifold ? there do exist examples on manifolds which are homeomorphic but not _ diffeomorphic _ to tori or infranilmanifolds @xcite . by putting conformality or `` pinching '' conditions on the restrictions of the derivative @xmath22 to @xmath63 and @xmath62 , researchers have established classification results in several settings ( see for example @xcite ) . the proof of [ thm : franksmanning ] considers the number of periodic points of a given period and relates this to the action of @xmath9 on homology groups @xmath136 using the lefschetz fixed point theorem . there has been some success on applying this technique for systems defined on other manifolds . for an @xmath137-dimensional sphere , the dimension of a homology group @xmath138 is either zero or one dimensional , and this may be used to show that no sphere supports an anosov diffeomorphism . using further techniques from algebraic topology , a. gogolev and f. rodriguez hertz have ruled out anosov diffeomorphisms on many families of manifolds , including certain sphere bundles and products of spheres with tori @xcite ( see also @xcite and references therein for more known obstructions ) . in the transitive case , other obstructions are obtained by studying the way the strong foliations `` wrap '' around the homology of the manifold via the study of _ currents _ ( see @xcite ) . the following important question remains open . is there a simply connected compact manifold admitting an anosov diffeomorphism ? spheres do not admit anosov diffeomorphisms because their homology is very small , but some of their products are already difficult to understand . in @xmath139 the proof that there are no anosov diffeomorphisms is very recent @xcite . it is unknown if @xmath140 admits anosov diffeomorphisms . this section discusses the work of m. brin , d. burago , and s. ivanov on the existence of branching foliations for 3-dimensional partially hyperbolic systems . as with anosov diffeomorphisms , invariant foliations play a fundamental role in the study of partially hyperbolic dynamics . we first present what could now be considered a classical result for partially hyperbolic diffeomorphisms @xcite . [ teo - strongunstablemanifold ] let @xmath141 be a partially hyperbolic diffeomorphism . then , there exist @xmath9-invariant foliations @xmath142 and @xmath143 tangent to the bundles @xmath144 and @xmath145 . moreover , the bundles @xmath144 and @xmath145 are uniquely integrable . if the stable bundle is not uniquely integrable at a point @xmath146 , then there are two distinct paths @xmath147 and @xmath148 starting at @xmath146 and tangent to @xmath62 . pick distinct points @xmath35 on @xmath147 and @xmath149 on @xmath148 near @xmath146 . as the stable bundle is contracting , @xmath150 is close to @xmath151 for all positive @xmath152 and there is a vector @xmath153 such that the exponential map based at the point @xmath151 satisfies @xmath154 . a sequence @xmath155 may be defined analogously and the same properties hold . we may assume that @xmath35 and @xmath149 were chosen so that the angle that @xmath156 makes with @xmath157 is very small . since @xmath158 contracts vectors in @xmath62 more sharply than those in @xmath72 and since @xmath159 is approximately equal to @xmath160 , one can show that the angle that @xmath161 makes with @xmath162 converges to zero as @xmath163 . the sequences @xmath164 and @xmath165 each tend to zero at a rate associated to the stable direction @xmath166 . this is strictly faster than the rate of contraction of @xmath167 which is associated to @xmath162 . this gives a contradiction . [ t ] . while this diagram does not commute , it comes arbitrarily close to commuting as @xmath163 and the distances between the points shrink to zero . the two triangles in the top left and top right represent small cones containing the @xmath72 direction . ] an important simple remark is the following : [ prop : wulines ] the foliations @xmath168 and @xmath169 have no compact leaves . we argue for @xmath168 . the argument for @xmath169 is symmetric . as in all of this section , we assume @xmath170 . suppose a compact leaf @xmath171 exists . then @xmath171 is a @xmath32 circle of some finite length . as @xmath158 contracts stable vectors , the length of @xmath172 shrinks to zero as @xmath173 . for large enough @xmath137 , @xmath172 is a circle lying inside a single foliation chart of @xmath168 , a contradiction . the main result of @xcite may be summarized as follows : [ thm : bi ] let @xmath0 be a partially hyperbolic diffeomorphism on a compact 3-manifold such that the bundles @xmath63 , @xmath61 , and @xmath62 are oriented and @xmath9 preserves these orientations . then , there is a collection @xmath69 of immersed surfaces tangent to @xmath71 such that every @xmath174 lies in at least one surface and no two surfaces in @xmath69 topologically cross . the collection is @xmath9-invariant : @xmath175 if and only if @xmath176 . the collection of surfaces @xmath69 in [ thm : bi ] is called a _ branching foliation _ and its elements are called leaves . distinct leaves may intersect without crossing , as depicted in [ fig : bran3 ] . [ t ] if @xmath9 satisfies the hypotheses of [ thm : bi ] then so does @xmath177 , and this means there is also a branching foliation @xmath70 tangent to @xmath72 . the conditions on orientation in the theorem do not hold in general . however , they may always be achieved by lifting an iterate of @xmath9 to a finite cover . throughout this survey paper , we lift to a finite cover and replace @xmath9 by an iterate @xmath178 whenever convenient . thus , for the remainder of the paper we always assume that the orientation conditions in [ thm : bi ] are satisfied . to be able to apply foliation theory , the following result from @xcite is also important . [ thm : separateleaves ] let @xmath179 and let @xmath69 be a branching foliation as in [ thm : bi ] . then there is a true foliation @xmath180 and a continuous surjective map @xmath181 with the following properties : 1 . every leaf of @xmath69 is the image @xmath182 of a leaf @xmath183 , 2 . the tangent plane @xmath184 is @xmath185-close to @xmath71 for all @xmath174 , and 3 . @xmath186 for all @xmath174 . for the interested reader , we spend the remainder of this section giving a rough outline of the ideas of the proofs of [ thm : bi , thm : separateleaves ] . in the three - dimensional setting , the center bundle is one - dimensional and therefore peano s existence result applies . this makes one wonder if it is possible to use the unique integrability of @xmath62 to obtain a foliation tangent to @xmath71 . however , even if @xmath61 were uniquely integrable , it may well be that @xmath71 is not integrable . this is the case for example for @xmath187 , which is rarely integrable even though each of @xmath62 and @xmath63 is uniquely integrable . using dynamical arguments , it is possible to show that a topological type of frobenius argument holds for the bundle @xmath71 . this is related to the fact that @xmath188 uniformly and strongly contracts vectors transverse to @xmath71 . [ prop : tansurf ] let @xmath189 be a curve tangent to @xmath71 and nowhere tangent to @xmath62 . then , the set @xmath190 is an immersed @xmath32 surface tangent to @xmath71 . this is a graph transform argument . for @xmath174 , let @xmath191 denote the stable curve of length one centered at @xmath35 . for each integer @xmath192 , let @xmath193 be a @xmath32 surface through @xmath194 which is approximately tangent to @xmath71 and which is @xmath195 close to @xmath196 . as @xmath197 strongly contracts vectors in @xmath63 , tangent planes to the surfaces @xmath198 converge uniformly to @xmath71 as @xmath173 . under the appropriate topology , the surfaces @xmath198 themselves converge to a surface tangent to @xmath71 and one can show that this limit surface is equal to @xmath190 . the main technical core of the paper @xcite deals with the construction of the collection of complete surfaces by making use of the previous proposition . using proposition [ prop : tansurf ] we know that given a center curve , one can saturate by strong stable manifolds in order to obtain a surface tangent to @xmath71 . notice that the size of such a surface for a given size of center curve is uniform . by this we mean that we can construct a uniform sized patch around each point of the center curve ( given by the fact that the bundles have uniformly bounded angles ) . the main difficulties are the following : 1 . one can not ensure that the saturation of a ( complete ) center curve is complete with the induced topology , 2 . if one considers all center curves one would get an @xmath9-invariant collection of surfaces , but they might topologically cross ; alternatively , one could consider only one center curve at each point so that the elements of the collection do not cross , but @xmath9-invariance becomes more difficult to establish . the proof uses the orientation of @xmath63 and the fact that @xmath158 preserves this orientation . this allows one to choose , for each point , its lowermost local surface tangent to @xmath71 and one will get both non - crossing and @xmath9-invariance . let us elaborate a bit on this . suppose @xmath199 is a small disk transverse to @xmath62 ( and thus also transverse to @xmath71 ) and with @xmath146 in the interior of the disk . let @xmath200 denote the intersection of @xmath71 with the tangent bundle @xmath201 . this intersection is a line field on the disk , and the orientation of @xmath61 gives an orientation to @xmath200 . with respect to this orientation , we may consider _ forward _ and _ backward _ curves tangent to @xmath200 . the orientation of @xmath63 gives a transverse orientation to @xmath71 which in turn gives a transverse orientation to @xmath200 . viewing @xmath200 as roughly horizontal , the transverse orientation gives a notion of `` up '' and `` down '' on the disk , and we may consider the _ lowest _ forward curve @xmath189 starting at @xmath146 and tangent to @xmath200 . by [ prop : tansurf ] , @xmath202 is an immersed surface tangent to @xmath71 . this surface may not be complete and it has a boundary consisting of stable leaves . consider now a point @xmath203 and a small immersed disk @xmath204 containing @xmath205 and transverse to @xmath62 . similar to before , define an oriented line field @xmath206 and let @xmath207 be the lowest forward curve tangent @xmath206 . if @xmath208 , there is nothing to do at @xmath205 . otherwise , define @xmath209 and glue @xmath210 to @xmath211 in order construct a larger surface tangent to @xmath71 . continuing in this way , one builds a surface @xmath212 tangent to @xmath71 such that @xmath213 consists of stable curves and the orientation of @xmath61 points into @xmath212 at all points on the boundary . the surface @xmath212 depends only on the point @xmath146 and the oriented bundles @xmath61 , @xmath62 , and @xmath63 . in particular , it does not depend on the various choices of disks and points @xmath199 , @xmath205 , @xmath204 , et cetera that occur in the construction . since @xmath61 points into @xmath212 , there is a uniform lower bound on the distance between points lying on distinct components of the boundary . if @xmath214 , then the surfaces @xmath212 and @xmath215 do not topologically cross this follows from the method of construction , though the proof is not immediate . since the partially hyperbolic diffeomorphism @xmath9 preserves the subbundles @xmath61 , @xmath62 , and @xmath63 and their orientations , it follows that @xmath216 [ p ! ] , take a center curve @xmath217 which is locally the lowest possible forward center curve through each of its points . let @xmath210 be the union of stable curves passing through @xmath217 ( a ) . at a point @xmath205 on the boundary of @xmath210 , the center bundle points away from @xmath210 and so one considers a lowest forward curve @xmath218 extending from @xmath205 . the stable curves through @xmath218 give a surface @xmath211 which is glued to @xmath210 ( b ) . by the same process , a similar surface is glued to the other boundary component of @xmath210 ( c ) . now in this example , every point on the boundary of the constructed surface @xmath212 points into the surface and this stage of the construction is finished ( d ) . the next stage will attach surfaces by considering center curves with the opposite orientation . ] now consider @xmath146 and @xmath199 as before and let @xmath219 be a backward curve starting at @xmath146 and tangent to @xmath200 . further , assume @xmath219 is the highest possible curve which does not topologically cross @xmath215 for any @xmath220 . attach the surface @xmath221 to @xmath212 . continuing in this way by attaching the stable saturates of highest backward curves , one builds a surface @xmath222 so that the orientation of @xmath61 now points away from @xmath223 at all points on the boundary . if @xmath214 , then @xmath223 and @xmath224 do not topologically cross . by alternating between forward and backward curves and at each step taking care not to cross a surface constructed in a previous step , one builds a nested sequence of surfaces @xmath225 which in the limit as @xmath163 yields a complete immersed surface through the point @xmath146 . we refer the reader to ( * ? ? ? * sections 46 ) for ( many ) more details . finally , let us mention that it is proven in ( * ? ? ? * lemma 7.1 ) that the collection can be chosen to satisfy that if @xmath226 and @xmath227 are surfaces of the collection through @xmath228 , then the sequence @xmath227 converges to a surface @xmath171 in the collection . we give here an indication of the proof of theorem [ thm : separateleaves ] . the idea is to flow leaves by a small amount depending on the point and the leaf in order to `` separate them '' . of course , it is essential that leaves do not cross in order that this will work . this can be done without problem in the core of a coordinate chart where we may consider a branching foliation on the unit cube @xmath229 ^ 3 \subset { \mathbb{r}}^3 $ ] where each leaf is the graph of a function @xmath229 ^ 2 \to { \mathbb{r}}$ ] . the condition that leaves do not topologically cross implies that if two leaves are given by the graphs of @xmath230 and @xmath231 , then one of the inequalities @xmath232 or @xmath233 holds pointwise for all @xmath234 ^ 2 $ ] . thus , the leaves are a totally ordered set . properties of ordered sets imply that we may associate to each leaf a number @xmath235 $ ] such that if @xmath236 is the leaf associated to @xmath237 , then @xmath238 now take @xmath239 and define functions @xmath240 this family of functions has the stronger property that @xmath241 the resulting graphs are therefore pairwise disjoint and form a true foliation . this solves the problem locally . to solve the problem globally , cover @xmath57 with a finite number of foliation charts and perform this procedure in each chart in parallel , using a partition of unity to make the result continuous and taking care to ensure that changes in one chart are small enough that they do not cause problems in other charts . see ( * ? ? ? * section 7 ) for details . the above sketch implicitly used the following property of branching foliations . [ prop : nobranch ] if the leaves of a branching foliation are pairwise disjoint , then the branching foliation is a true foliation . see the discussion in ( * ? ? ? * remark 1.10 ) for further details . for the classification results in this survey , this property will allow us to establish dynamical coherence as a step towards leaf conjugacy . integrability of @xmath242 and @xmath243 into @xmath9-invariant foliations is called _ dynamical coherence _ as introduced in the first section of this survey ( see also @xcite ) . it is unknown in general if transitive partially hyperbolic diffeomorphisms in dimension 3 are dynamically coherent . there exist non - transitive examples of non - dynamically coherent partially hyperbolic diffeomorphisms @xcite . we refer the reader to @xcite for a survey on this problem in 3-dimensional manifolds . in this survey , integrability of the bundles appears as a step in the quest for showing leaf conjugacy to the known models . as a consequence of proposition [ prop : nobranch ] we have : for a partially hyperbolic diffeomorphism @xmath9 on a compact 3-manifold , if the leaves of the branching foliation @xmath69 are pairwise disjoint and the leaves of the branching foliation @xmath70 are pairwise disjoint , then @xmath9 is dynamically coherent . in the case of a strongly partially hyperbolic system , the tangent space @xmath17 splits into three subbundles , each of dimension at least one . the manifold @xmath57 must therefore have dimension at least three . for a long time , this was the only known obstruction ; it was an open question whether every manifold @xmath57 with @xmath244 supported a partially hyperbolic diffeomorphism . this was finally answered by brin , burago , and ivanov , who used the branching foliations described in [ sec : integrability ] to show the following . there are no partially hyperbolic diffeomorphisms on the 3-sphere . the proof relies on the theory of compact leaves and reeb components developed by novikov @xcite . we give a brief overview of this theory and refer the reader to @xcite for a detailed treatment . first , consider an annulus @xmath245 $ ] constructed by identifying the left and right sides of a square . on this annulus , consider a foliation as in [ fig : reeba ] with two circle leaves on the boundary and where all other leaves are lines . [ t ] finally , rotate @xmath245 $ ] around the axis @xmath246 to construct a solid torus @xmath247 . the resulting two dimensional foliation has the boundary as a compact leaf and all other leaves are topological planes . the solid torus equipped with this foliation is called a reeb component . suppose @xmath57 is a 3-manifold and @xmath37 a two - dimensional foliation on @xmath57 . if there is a solid torus @xmath248 such that the foliation @xmath37 restricted to @xmath249 is equivalent to the construction above , then @xmath249 is a reeb component for @xmath37 . the embedding @xmath250 induces a homomorphism @xmath251 . since the embedding factors as a composition @xmath252 , it is easy to see that @xmath253 is not injective . novikov proved the much more difficult converse . [ thm : piinj ] suppose @xmath37 is a two - dimensional foliation of a 3-manifold . if there is a leaf @xmath171 such that @xmath254 is not injective , then @xmath37 has a reeb component . this was originally proved for smooth foliations and later extended to the @xmath195 case by solodov @xcite . call a foliation _ reebless _ if it has no reeb components . the above results then say that a foliation is reebless if and only if every leaf is @xmath255-injectively immersed in @xmath57 . [ prop : reebcircle ] suppose @xmath37 is a two - dimensional foliation and @xmath256 a transverse one - dimensional foliation on a 3-manifold . if @xmath37 has a reeb component , then @xmath256 has a compact leaf . assume @xmath256 is orientable , so that there is a flow whose orbits are exactly the leaves of @xmath256 . suppose @xmath37 has a reeb component @xmath249 and assume on the boundary of @xmath249 that the direction of the flow is pointing into the solid torus . then any orbit which enters @xmath249 can not leave @xmath249 . with some work , one can show that there a compact disk lying inside a leaf in @xmath249 to which every orbit of the flow returns . the brouwer fixed point theorem then implies the flow has a periodic orbit . novikov also established a form of converse to [ prop : reebcircle ] . given a foliation @xmath37 , call a circle @xmath257 a _ transverse contractible cycle _ if is it contractible ( that is , homotopic in @xmath57 to a constant path ) and transverse to @xmath37 . ( while the original @xmath210 is transverse to @xmath37 , the circles in the homotopy need not be . ) [ prop : transversecycle ] a reebless foliation has no transverse contractible cycle . for the remainder of the section , we consider foliations for the specific case of a partially hyperbolic diffeomorphism @xmath9 on a 3-manifold @xmath57 . let @xmath169 be the unstable foliation . recall that proposition [ prop : wulines ] implies that @xmath169 has no circle leaves . by [ thm : bi , thm : separateleaves ] , assume @xmath69 is a branching foliation approximated by a true foliation @xmath180 . take @xmath185 small enough that @xmath180 is transverse to @xmath169 . and [ prop : wulines ] immediately give the following . the foliation @xmath180 is reebless . let @xmath258 be the universal cover of @xmath57 . as @xmath180 is a true foliation , lifting it to a foliation @xmath259 on @xmath258 is straightforward . to lift the branching foliation , let @xmath181 be the map which takes leaves of @xmath180 to leaves of @xmath69 . lift @xmath260 to @xmath261 . then the leaves of @xmath262 are exactly the surfaces of the form @xmath263 where @xmath264 . every leaf of @xmath259 is a plane . suppose @xmath265 is a leaf of @xmath259 and @xmath266 is a closed path in @xmath265 . let @xmath171 and @xmath189 be their projections in @xmath57 . since @xmath266 is contractible in @xmath258 , @xmath189 is contractible in @xmath57 . by [ thm : piinj ] , @xmath189 is also contractible in @xmath171 . the homotopy in @xmath171 taking @xmath189 to a point lifts to a homotopy in @xmath265 taking @xmath266 to a point . this shows that @xmath265 is simply connected . the surface @xmath265 is then either a plane or a 2-sphere . the map @xmath267 restricts to a homeomorphism between @xmath265 and a leaf of @xmath180 . this leaf @xmath263 has a splitting @xmath268 of its tangent bundle and therefore can not be a sphere . palmeira @xcite showed that if a simply connected @xmath269-manifold @xmath258 is foliated by planes of dimension @xmath270 , then @xmath258 is homeomorphic to @xmath271 . thus in the current setting , @xmath258 must be homeomorphic to @xmath272 . due to its importance , we state this as a self - contained theorem . if a 3-manifold supports a partially hyperbolic diffeomorphism , its universal cover is homeomorphic to @xmath272 . though @xmath69 may have branching , in many ways it behaves like a reebless foliation . there is no contractible cycle transverse to @xmath69 . suppose @xmath189 is a transverse cycle . the angle between @xmath189 and @xmath69 is bounded away from zero , say by @xmath239 . take the approximation @xmath180 with @xmath273 . then @xmath189 is also transverse to @xmath69 and therefore not contractible . this implies a uniqueness property for intersections on the universal cover . [ prop : uscunique ] a leaf of @xmath262 intersects a leaf of @xmath274 in at most one point . suppose @xmath275 and that @xmath276 intersects @xmath171 at a point @xmath277 . a new path can be constructed by following @xmath276 from @xmath35 to a point @xmath278 just before @xmath149 , and then connecting @xmath278 to @xmath35 by a path which stays close to @xmath171 and which is everywhere transverse to @xmath171 . [ t ] in distinct points @xmath35 and @xmath149 , a transverse cycle may be constructed . ] see [ fig : transverse ] . the result is a cycle in @xmath258 transverse to @xmath262 . this projects down to a transverse contractible cycle in @xmath57 . [ cor : halfspace ] each leaf @xmath171 of @xmath262 is a properly embedded plane which splits @xmath258 into two half spaces . if @xmath171 is not properly embedded , it accumulates at a point @xmath279 and intersects the unstable leaf through @xmath35 in infinitely many points . implies that different segments of the same unstable leaf must stay a certain distance away from each other . this notion may be stated quantitatively . for a subset @xmath280 , let @xmath281 be the neighborhood of all points at distance at most @xmath282 from @xmath119 . that is , @xmath283 if @xmath284 . [ lem : volume ] there is @xmath285 such that if @xmath286 is an unstable curve , then @xmath287 for an unstable segment @xmath288 , define @xmath289 as those points @xmath290 such that @xmath291 for every leaf @xmath292 with @xmath293 . since @xmath71 and @xmath63 are transverse , one can find @xmath239 such that @xmath294 for all unstable curves @xmath288 of length exactly one . if @xmath295 for some integer @xmath137 , there are pairwise disjoint subcurves @xmath296 each of length one . by [ prop : uscunique ] , the sets @xmath297 are pairwise disjoint which implies that @xmath298 . if @xmath286 is an unstable curve , then the length @xmath299 grows exponentially fast . if the universal cover @xmath258 is `` small '' in some geometric sense , then this is only possible if the action @xmath80 on @xmath300 exhibits some form of partial hyperbolicity . to give an idea of this , we prove a simple result on @xmath75 . if @xmath301 is partially hyperbolic , then @xmath302 is not the identity . suppose @xmath80 is the identity . then , @xmath9 lifts to a map @xmath303 which is a finite distance from the identity map on @xmath272 . there is @xmath304 such that if @xmath305 denotes the ball of radius @xmath306 , then @xmath307 for all @xmath308 . take an unstable segment @xmath309 . then , @xmath310 which grows at most polynomially in @xmath137 . since @xmath311 grows exponentially , [ lem : volume ] gives a contradiction . since @xmath312 , the homomorphism @xmath80 may be regarded as a @xmath313 matrix with eigenvalues @xmath314 ordered by @xmath315 . then @xmath316 by gelfand s formula and the above proof can be adapted to show that @xmath317 . a similar study of @xmath177 shows that @xmath318 . this is explored in more detail in [ sec : skew , sec : da ] . the arguments in the previous section all involved the one - dimensional unstable foliation @xmath169 and the two - dimensional branching foliation @xmath69 . once we established the existence of the branching foliation @xmath69 , the finer splitting @xmath319 was not directly used . one could therefore ask if the arguments also apply to weakly partially hyperbolic systems . indeed , the following holds . [ prop : weakfoln ] let @xmath0 be a weakly partially hyperbolic diffeomorphism with a splitting of the form @xmath320 where @xmath321 and @xmath133 . further , suppose there is a foliation @xmath37 transverse to @xmath63 . then : 1 . the universal cover @xmath258 is @xmath272 , 2 . the leaves of the lifted foliation @xmath322 are all planes , and 3 . each leaf of @xmath274 intersects each leaf of @xmath322 at most once . this was used in @xcite to study dynamical coherence of weakly partially hyperbolic diffeomorphisms isotopic to a linear anosov automorphism of @xmath323 ( and extended in @xcite to higher dimensions ) . the notion of _ almost dynamical coherence _ was used there to refer to the existence of a foliation @xmath37 transverse to @xmath63 as in the proposition . the problem is that in the weakly partially hyperbolic setting , we have no tools at present to find such a foliation @xmath37 . let @xmath0 be a weakly partially hyperbolic diffeomorphism with a splitting of the form @xmath320 where @xmath321 and @xmath133 . is there a branching foliation tangent to @xmath61 ? is there a true foliation transverse to @xmath63 ? it is not clear even locally if there exist small surfaces tangent to @xmath71 . some progress in this direction has been obtained by luzzatto , turelli and war @xcite continuing the work in @xcite and recently improvements have been announced by turelli and war . if there always exists a transverse foliation @xmath37 as in [ prop : weakfoln ] , it would imply that any 3-manifold which supports a robustly transitive diffeomorphism must also support a reebless foliation . that would definitively answer the following long standing open question . does @xmath324 support a robustly transitive diffeomorphism ? in the measure - preserving setting , one can also ask if @xmath324 supports a stably ergodic diffeomorphism , a question which is also wide open . one can consider even weaker forms of @xmath158-invariant structures and make an attempt at classification . one such invariant structure occurs for _ non - uniformly hyperbolic _ systems . in this case , the splitting of the tangent bundle is only measurable instead of continuous . therefore it does not impose much restrictions on the topology of the underlying manifolds , and indeed , it is shown in @xcite that every compact manifold admits a conservative non - uniformly hyperbolic diffeomorphism . another case is a _ dominated splitting _ where the tangent bundle has a @xmath158-invariant splitting @xmath325 such that @xmath326 for some @xmath327 and all @xmath174 and unit vectors @xmath328 and @xmath329 . in comparison to weak partial hyperbolicity , a dominating splitting puts no conditions of uniform expansion or contraction on either @xmath330 or @xmath113 . gourmelon and potrie are currently preparing a complete classification of the dynamics and structure of diffeomorphisms of surfaces admitting global dominated splittings @xcite which builds on the work of pujals and sambarino who have studied in great detail the dynamics of surface diffeomorphisms whose limit set admits a dominated splitting @xcite . in the partially hyperbolic context , we mostly consider invariant foliations without compact leaves . this is due to the following work of rodriguez hertz , rodriguez hertz , and ures @xcite . [ thm : anosovtorus ] suppose @xmath9 is a partially hyperbolic diffeomorphism of a compact 3-manifold @xmath57 and there is a compact surface tangent to @xmath71 . then : 1 . there is a periodic 2-torus @xmath331 tangent to @xmath71 , 2 . the map @xmath332 is isotopic to an anosov diffeomorphism on @xmath333 , and 3 . ( up to finite cover ) @xmath57 either is the 3-torus or the suspension of an anosov diffeomorphism . the torus @xmath333 is a specific case of an _ anosov torus _ , a concept covered in detail in @xcite and so we do not focus on its properties here . foliations @xmath334 and @xmath335 on a manifold @xmath57 are _ equivalent _ if there is a homeomorphism @xmath336 such that @xmath171 is a leaf of @xmath334 if and only if @xmath337 is a leaf of @xmath335 . when the foliations are @xmath6 smooth for @xmath338 , equivalence is a natural notion to use . for instance , the following holds for foliations on surfaces . [ prop : c2foln ] let @xmath256 be a @xmath339 foliation without compact leaves defined on @xmath129 . then , @xmath256 is equivalent to a foliation by lines of constant irrational slope . if @xmath256 is only @xmath32 , the result does not hold , due to foliations with exceptional minimal sets as discovered by denjoy . for partially hyperbolic systems , the foliations under study are typically @xmath195 foliations with @xmath32 leaves . to analyse these foliations , it has turned out to be useful to consider a weaker equivalence of foliations without compact leaves @xcite . two ( branching ) foliations @xmath334 and @xmath335 of @xmath258 are _ almost parallel _ if there exists @xmath340 such that : * for every @xmath341 there exists @xmath342 such that @xmath343 , * for every @xmath342 there exists @xmath341 such that @xmath343 . here , @xmath344 denotes the hausdorff distance between closed subsets of @xmath258 with a metric obtained by lifting the metric of @xmath57 to the universal cover . note that `` almost parallel '' is an equivalence relation , but the word `` equivalent '' is reserved for the stronger notion defined above . [ prop : c0foln ] let @xmath256 be a @xmath195 foliation without compact leaves defined on @xmath129 . then , @xmath256 is almost parallel to a foliation by lines of constant irrational slope . we leave this as an exercise to the reader . plante classified the possible @xmath339 reebless foliations in 3-manifolds with solvable fundamental group @xcite . adapting parts of the proofs , we established a weaker result for @xmath195 branching foliations ( * ? ? ? * appendix b ) . see also @xcite . to state the result , we introduce the notion of a model foliation . first , consider a linear two - dimensional foliation on @xmath272 . that is , a foliation by geometric planes . restrict this foliation to @xmath345 $ ] and quotient by the lattice @xmath346 to produce a linear foliation @xmath37 on @xmath347 $ ] . let @xmath348 and @xmath334 denote the restrictions of @xmath37 to the boundary components @xmath349 and @xmath350 now consider a linear map @xmath351 and identify @xmath352 with @xmath353 to produce the suspension manifold @xmath354 . if @xmath355 maps @xmath348 to @xmath334 , then the linear foliation @xmath37 yields a foliation @xmath356 on the @xmath354 . we call such a foliation @xmath356 a _ model foliation_. in the special case that @xmath355 is the identity , @xmath357 and the model foliations are precisely the linear foliations . [ thm : models ] suppose that @xmath37 is a ( branching ) foliation on a compact 3-manifold with solvable fundamental group and that @xmath37 has no compact leaves . then , ( up to finite cover ) @xmath57 is a torus bundle over a circle and @xmath37 is almost parallel to a model foliation . some classification results also hold for reebless foliations in other families of 3-manifolds . see the books @xcite for comprehensive treatments of foliation theory in dimension 3 and beyond . states some results for reebless foliations on seifert spaces and applies these to the study of partial hyperbolicity on such manifolds . we give a broad outline of how to establish a leaf conjugacy such that all of the results for the 3-torus and other nilmanifolds and solvmanifolds follow roughly the same blueprint . the fact that these different classification results can be fit into the same framework gives hope that some of the techniques will also apply to other partially hyperbolic systems . given a partially hyperbolic diffeomorphism @xmath9 on a 3-manifold @xmath57 , the first step is to use properties of the geometry of @xmath57 to characterize the action @xmath358 on the fundamental group . from this , construct a diffeomorphism @xmath359 which is algebraic and such that @xmath360 . then , use the information about @xmath80 to show that @xmath14 is partially hyperbolic . since the universal cover @xmath258 is homeomorphic to @xmath272 , results in algebraic topology imply that @xmath9 and @xmath14 are homotopic ( * ? ? ? * proposition 1b.9 ) . lift @xmath9 and @xmath14 to diffeomorphisms @xmath106 and @xmath361 which are at finite distance from each other ; that is , @xmath362 implies that there is a reebless branching foliation @xmath69 tangent to @xmath71 on @xmath57 and invariant by @xmath9 . lift @xmath69 to a branching foliation @xmath262 on @xmath258 . by [ thm : models ] , there is a model foliation @xmath363 with lift @xmath364 such that @xmath262 is almost parallel to @xmath364 . now consider a leaf @xmath365 . it lies a finite distance from a leaf @xmath292 . then @xmath366 lies a finite distance from @xmath367 . using the definition of almost parallel , this implies that @xmath367 is a finite distance from a leaf of @xmath364 . in other words , @xmath368 is almost parallel to @xmath364 . this further restricts the possibilities for @xmath364 . using techniques specific to the manifold , show that @xmath364 is equal to the center - stable foliation @xmath369 associated to @xmath361 . this establishes that @xmath262 is almost parallel to @xmath369 . similarly , show that @xmath370 is almost parallel to the center - unstable foliation @xmath371 of @xmath361 . our final goal is to construct a homeomorphism @xmath336 homotopic to the identity which is a leaf conjugacy between @xmath9 and @xmath14 . note that if such a map @xmath94 exists , it lifts to a map @xmath372 which is a finite distance from the identity . consider @xmath373 and let @xmath374 be the leaf which passes through @xmath375 . the fact that @xmath376 is a leaf conjugacy implies that @xmath377 for all @xmath378 . then , @xmath379 since we do not have the leaf conjugacy @xmath94 in advance , we run this reasoning backwards . that is , we establish that for every @xmath373 there is exactly one leaf @xmath374 which satisfies @xmath380 . this then defines a map @xmath381 by setting @xmath382 equal to @xmath116 . one must further show that the leaves of @xmath262 are the fibers of @xmath383 . since the fibers of @xmath383 are disjoint , this implies that @xmath262 is a true foliation . further , it is the unique invariant foliation , since the definition of @xmath383 does not depend on the choice of @xmath262 . similar reasoning gives a map @xmath384 . assuming each leaf of @xmath369 intersects each leaf of @xmath371 in exactly one center leaf , these two maps combine to give @xmath385 where @xmath386 . thus , each point of @xmath35 is mapped to a center leaf @xmath387 of @xmath361 in such a way that @xmath388 is constant on center leaves of @xmath106 . the final step is to define a homeomorphism @xmath372 such that @xmath389 for all @xmath279 and in such a way that @xmath376 quotients down to a homeomorphism @xmath336 on the original compact manifold . this @xmath94 will be the desired leaf conjugacy . the original papers have three very different methods for constructing @xmath376 from @xmath388 . with the goal of unifying these three proofs as much as possible , we show that the averaging technique used in @xcite can be used in all of the cases . this averaging method is due to fuller @xcite , who used it to study cross sections of flows . others have employed the same technique @xcite and it appears to have been independently discovered multiple times . the next step uses a smooth projection @xmath390 with the following properties . 1 . for any @xmath391 , @xmath392 is a @xmath32 diffeomorphism . 2 . there is @xmath393 , such that if @xmath35 and @xmath149 are on the same center leaf @xmath171 of @xmath106 and their distance @xmath394 measured along @xmath171 is exactly @xmath333 , then @xmath395 . 3 . there is a homomorphism @xmath396 such that @xmath397 for all @xmath279 and @xmath398 . the choice of @xmath399 depends on the @xmath400-manifold @xmath57 . now consider @xmath401 and let @xmath402 be a function which parameterizes @xmath171 by arc length . for @xmath293 , let @xmath403 and define @xmath404 by the condition @xmath405 this defines a homeomorphism from @xmath171 to @xmath406 which is independent of the choice of @xmath147 . when the same construction is performed for every leaf in @xmath407 the result is a leaf conjugacy @xmath372 . the existence of the homomorphism @xmath408 implies that @xmath376 commutes with every deck transformation , and therefore descends to a leaf conjugacy on the compact manifold @xmath57 . the next four sections discuss steps to establish leaf conjugacy which are unique to the manifolds under study . let @xmath301 be partially hyperbolic . then @xmath9 lifts to @xmath303 and @xmath106 is a finite distance from a linear map @xmath409 . let @xmath314 be the eigenvalues of @xmath100 ordered so that @xmath315 . in this section , we consider the case @xmath410 and give an outline of the proof of [ thm : nil ] in this setting . replacing @xmath9 by an iterate if necessary , we assume that all of the eigenvalues are positive . as noted at the end of [ sec : reebless ] , the eigenvalues satisfy @xmath411 . if @xmath37 is a reebless branching foliation in @xmath75 , then @xmath322 is almost parallel to a linear foliation , that is , the translates of a plane . if @xmath412 is almost parallel to @xmath322 , then this plane is invariant and is therefore the span of two eigenspaces of @xmath100 . denote the possible linear foliations as @xmath413 , @xmath371 , and @xmath369 corresponding to the partially hyperbolic splitting of @xmath100 . we consider the three possibilities in turn . [ prop : ustorus ] if @xmath262 is almost parallel to @xmath413 , then @xmath69 contains a compact leaf . since [ thm : nil ] assumes no compact leaves , this eliminates one of the possibilities . since @xmath414 , there is a rank-2 subgroup @xmath415 corresponding to @xmath416 . take any @xmath292 . by [ cor : halfspace ] , @xmath171 splits @xmath272 into two half spaces @xmath417 and @xmath418 . consider the union @xmath419 . since @xmath262 is almost parallel to @xmath413 , this union is not the entire space @xmath272 and therefore has a non - empty boundary . each component of this boundary is a leaf of @xmath262 which is invariant under @xmath420 and therefore quotients down to a 2-torus embedded in @xmath75 . [ lem : boxgrow ] @xmath370 is not almost parallel to @xmath369 . consider a coordinate system on @xmath272 such that @xmath100 is given by @xmath421 this change of coordinates affects volumes and distances in @xmath272 by at most a constant factor . consider the cube @xmath422 \times [ -1,1 ] \times [ -1,1 ] . $ ] as @xmath106 is a bounded distance from @xmath100 , one can show that there is @xmath423 such that @xmath424 \times [ -({\lambda}_3+c)^n , ( { \lambda}_3+c)^n ] \times [ -c n , c n].\ ] ] for all @xmath425 . now suppose @xmath288 is an unstable curve inside @xmath119 . if @xmath370 is almost parallel to @xmath369 , then each @xmath299 lies in an @xmath249-neighbourhood of a leaf of @xmath369 . this means there is a sequence @xmath426 such that @xmath427 \times [ y_n - r , y_n+r ] \times [ -c n , c n].\ ] ] as the volume of this box grows only linearly in @xmath137 , [ lem : volume ] gives a contradiction . this last proof is a specific case of the idea of confining an unstable segment of @xmath106 to the @xmath249-neighbourhood of a leaf of a model foliation and using [ lem : volume ] to derive a contradiction . this technique occurs many times in the proofs of the classification results and we refer to it as the `` length versus volume '' argument . analogously , @xmath262 is not almost parallel to @xmath371 and therefore @xmath262 is almost parallel to @xmath369 . by employing the semiconjugacy of franks ( [ thm : franks ] ) , one can show that there is a map @xmath428 defined by the property that @xmath382 is the unique leaf of @xmath369 for which @xmath429 this map is continuous , surjective , and commutes with deck transformations . further , the supremum above is bounded by a constant @xmath210 independent of @xmath146 . the most important property of @xmath383 is that its fibers are exactly the leaves of @xmath262 . to see this , first suppose points @xmath146 and @xmath430 lie on the same leaf @xmath292 . by the definition of almost parallel , @xmath431 lies in the @xmath249-neighbourhood of a leaf @xmath432 and therefore @xmath433 the same inequality holds with @xmath430 in place of @xmath146 and so @xmath434 then @xmath435 which in the limit @xmath173 implies @xmath436 . now suppose distinct leaves @xmath437 are both mapped by @xmath383 to @xmath374 . a priori , @xmath171 and @xmath438 could be disjoint or , in the case of a branching foliation , could intersect without being equal . either way , the region between @xmath171 and @xmath438 has non - empty interior @xmath439 and @xmath440 for every point @xmath441 . let @xmath288 be an unstable curve in @xmath439 . every point in @xmath299 is mapped to @xmath442 . a minor adaptation of the proof of [ lem : boxgrow ] then gives a contradiction . thus the fibers of @xmath383 are disjoint surfaces and form a true , non - branching foliation tangent to @xmath443 . since the definition of @xmath383 did not depend on @xmath69 , this invariant foliation is unique . a similar construction holds for @xmath444 the product @xmath445 yields a homeomorphism between the space of center leaves of @xmath106 and those of @xmath100 . let @xmath446 be a linear map transverse to the center direction of @xmath100 . using this @xmath399 , the averaging method of fuller described in [ sec : leafconj ] gives the leaf conjugacy . let @xmath301 be partially hyperbolic . then @xmath9 lifts to @xmath303 and @xmath106 is a finite distance from a linear map @xmath409 . let @xmath314 be the eigenvalues of @xmath100 ordered so that @xmath315 . in contrast to [ sec : skew ] , we now discuss the proof of [ thm : nil ] for the case @xmath447 . up to replacing @xmath9 by its inverse , assume @xmath448 [ lem : real ] the eigenvalues @xmath314 are real and distinct . assume @xmath449 and @xmath450 are complex . then there is a unique @xmath100-invariant two - dimensional subspace @xmath451 and @xmath452 is a contraction . by this uniqueness , @xmath262 is almost parallel to the translates of @xmath399 . one can then consider a stable arc @xmath288 and use the contraction of @xmath452 in a length versus volume argument to get a contradiction . thus , the @xmath314 are real . algebraic properties of @xmath453 imply that the eigenvalues are distinct ( see for example ( * ? ? ? * definition v.3.10 ) and the remark that follows ) . has irrational eigenvalues @xmath454 counted with multiplicity . irrationality implies that @xmath455 is the minimal polynomial for the @xmath314 . using @xmath456 and @xmath457 where @xmath458 is the trace of @xmath100 , observe that @xmath459 . uniqueness of the minimal polynomial implies that @xmath460 which is impossible for a matrix with integer entries . thanks goes to m. coons for aiding in this proof . ] there are exactly three linear foliations with property that @xmath461 is almost parallel to @xmath364 . these are the foliations @xmath369 , @xmath371 , and @xmath413 corresponding to the partially hyperbolic splitting . descending to the compact manifold , @xmath462 , @xmath463 , and @xmath464 are foliations of @xmath75 by planes . [ prop : planes ] let @xmath37 be a two - dimensional ( branching ) foliation on @xmath75 and @xmath256 a transverse one - dimensional foliation . suppose @xmath363 is a linear foliation by planes such that the lifts @xmath322 and @xmath364 are almost parallel . then 1 . @xmath37 is a ( branching ) foliation by leaves diffeomorphic to planes ; 2 . the lifted foliations @xmath322 and @xmath465 have global product structure ; that is , each leaf of @xmath322 intersects each leaf of @xmath465 in exactly one point ; and 3 . there are @xmath466 and a closed cone @xmath467 such that @xmath468 for any points @xmath146 and @xmath430 lying on the same leaf of @xmath465 . moreover , any line @xmath469 is transverse to @xmath364 . if @xmath37 were not a foliation by planes , there would be @xmath470 and @xmath471 such that @xmath472 . one can then show that @xmath171 does not lie in an @xmath249-neighbourhood of any @xmath365 . if every leaf of a foliation is simply connected , the foliation is without holonomy and global product structure follows from results in foliation theory ( see @xcite ) . the third item can be proved using global product structure and the definition of almost parallel . let @xmath473 be the lifted semiconjugacy of franks ( ) . it is a finite distance from the identity and satisfies @xmath474 . @xmath370 is almost parallel to @xmath371 . let @xmath364 be the linear foliation almost parallel to @xmath370 . note that @xmath475 where @xmath476 is the eigenspace associated to @xmath477 . by [ prop : planes ] , @xmath476 is contained in a cone @xmath478 transverse to @xmath364 . hence , @xmath479 as that is the only linear @xmath100-invariant foliation transverse to @xmath476 . [ lem : dapar ] @xmath262 is almost parallel to @xmath369 . first , suppose @xmath262 is almost parallel to @xmath371 . then , the same argument as in the proof of [ lem : real ] gives a contradiction next , suppose @xmath262 is almost parallel to @xmath413 and let @xmath480 be the cone transverse to @xmath413 from [ prop : planes ] . if @xmath481 , then @xmath482 for large positive @xmath137 . then , @xmath483 is bounded away from @xmath484 as @xmath173 . this is only possible if @xmath485 . the previous proof showed that if @xmath486 , then @xmath487 . thus for any leaf @xmath374 , @xmath488 is saturated by both the stable and unstable leaves of @xmath106 . this is a non - generic property . by perturbing @xmath9 in the space of @xmath32 diffeomorphisms , we may assume that @xmath9 is _ accessible _ and consequently that @xmath489 is the only non - empty subset @xmath490 which is saturated by both stable and unstable leaves of @xmath106 . one can show that the perturbation does not modify the asymptotic directions of the foliations . this gives a contradiction . as @xmath371 and @xmath413 have been eliminated , the only remaining possibility is that @xmath262 is almost parallel to @xmath369 . using [ lem : dapar ] and the properties of @xmath388 , one can show that for any @xmath292 the image @xmath406 is equal as a subset of @xmath272 to a leaf of @xmath369 . suppose distinct leaves @xmath437 satisfy @xmath491 . by the global product structure given by [ prop : planes ] , there is an unstable curve @xmath288 with endpoints @xmath492 and @xmath493 . for large positive @xmath137 , @xmath494 where @xmath495 is a cone transverse to @xmath369 given by [ prop : planes ] . therefore , @xmath496 is bounded away from @xmath497 as @xmath173 giving a contradiction . this shows that @xmath388 defines a bijection between the leaves of @xmath262 and those of @xmath369 . a similar result holds for @xmath370 and the remaining steps for establishing the leaf conjugacy are exactly as for the systems considered in [ sec : skew ] . we next consider the proof of [ thm : nil ] for 3-manifolds with nil geometry . these are non - trivial circle bundles over @xmath129 . as the bundle is non - trivial , there can not be a horizontal surface . because of this property , all leaves of a reebless foliation are vertical . in some ways , this makes these manifolds simpler to study than @xmath75 even though the geometry is more difficult to visualize . to give a precise description of the reebless foliations , we first look at the universal cover . define heisenberg space @xmath498 as the nilpotent lie group consisting of matrices of the form @xmath499 in this section , we write the above element as @xmath500 for brevity . every compact 3-manifold with nil geometry can be defined as the quotient of @xmath498 by a discrete subgroup . let @xmath501 be such a manifold . let @xmath502 be of the form @xmath503 for some non - zero @xmath504 . the level sets of @xmath505 define a foliation @xmath506 of @xmath498 by planes which quotients down to a foliation @xmath507 on @xmath501 . if @xmath37 is a reebless foliation on @xmath501 , its lift @xmath322 is almost parallel to @xmath506 for some @xmath505 . this is a restatement of [ thm : models ] in this specific setting . now suppose @xmath508 is partially hyperbolic . since @xmath501 is a nilmanifold , the results of malcev @xcite imply that there is a lie group automorphism @xmath509 such that @xmath510 quotients down to a diffeomorphism @xmath511 homotopic to @xmath9 . this automorphism is of the form @xmath512 where @xmath513 is a linear map with @xmath514 and @xmath146 is a quadratic polynomial . a length versus volume argument implies that @xmath100 has an eigenvalue @xmath515 and so @xmath14 is partially hyperbolic . let @xmath369 and @xmath371 be the invariant foliations of @xmath510 . since @xmath516 is not integrable , there is no @xmath413 foliation . for a one dimensional subspace @xmath517 , the following properties are equivalent : * @xmath330 lies at a finite distance from @xmath518 , * @xmath330 is an eigenspace of @xmath100 . from this , one can show that if @xmath506 is almost parallel to @xmath519 , then either @xmath520 or @xmath521 . if @xmath370 is almost parallel to @xmath369 , a variant of the proof of [ lem : boxgrow ] gives a contradiction . therefore , @xmath370 is almost parallel to @xmath371 and similarly @xmath262 is almost parallel to @xmath369 . from this point on , the proof is as in [ sec : skew ] with minor variations . for the averaging method of fuller , use @xmath522 . this section discusses the proof of [ thm : sol ] . let @xmath523 be a hyperbolic toral automorphism and define the suspension flow @xmath355 on the manifold @xmath524 . suppose @xmath9 is a partially hyperbolic diffeomorphism on @xmath525 . by replacing @xmath9 with an iterate , we may assume @xmath9 is homotopic to the identity . then @xmath9 lifts to @xmath106 on the universal cover @xmath526 where @xmath106 is a finite distance from the identity . the universal cover @xmath258 is homeomorphic to @xmath527 , and so we will at times write points in @xmath258 in the form @xmath528 , but keep in mind that the geometry is not euclidean . the anosov flow @xmath355 lifts to a flow @xmath529 given by @xmath530 . the time - one map is partially hyperbolic . let @xmath369 , @xmath371 , and @xmath413 be the corresponding model foliations . any reebless foliation @xmath37 on @xmath525 lifts to a foliation @xmath322 on @xmath526 which is almost parallel to one of @xmath369 , @xmath371 , and @xmath413 . the following result on compact leaves is proved in a similar way to [ prop : ustorus ] . if @xmath37 lifts to a foliation almost parallel to @xmath413 , then @xmath37 has a compact leaf . as we are under the hypothesis that @xmath262 has no torus leaves , it follows that @xmath262 is almost parallel either to @xmath369 or @xmath371 . up to possibly replacing @xmath100 by @xmath531 , we may freely assume that @xmath262 is almost parallel to @xmath369 . our goal is then to show that @xmath370 is almost parallel to @xmath371 . [ lem : cuap ] @xmath370 is almost parallel to @xmath371 . this step was surprisingly difficult . it was the last piece of the puzzle when the results in @xcite were developed and frustrated the authors for some time . sadly , its proof is not easily summarized and so for this survey we just take [ lem : cuap ] as a given . for every @xmath292 , there is a unique @xmath374 such that @xmath171 and @xmath116 are at finite distance . by the definition of almost parallel , at least one such leaf exists . since the elements of @xmath369 are the weak stable leaves of a lifted anosov flow , no two of them are at finite distance . there is also a converse result , swapping the roles of @xmath262 and @xmath369 . for every @xmath374 , there is a unique @xmath292 such that @xmath171 and @xmath116 are at finite distance . suppose two leaves @xmath532 are both at finite distance from @xmath374 . then one can take an unstable curve @xmath288 lying between @xmath171 and @xmath438 and apply a length versus volume argument for @xmath299 to derive a contradiction . in this case , the details are complicated and one must consider volume on an intermediate covering space homeomorphic to @xmath533 . @xmath534 for all @xmath292 . since @xmath106 is a finite distance from the identity , both @xmath171 and @xmath535 are a finite distance from the same leaf @xmath374 . since we are trying to compare the system to a flow , knowing that @xmath106 fixes leaves is an important step . paradoxically , fixing leaves implies that @xmath106 does not fix points . this was first observed in @xcite . [ lem : noper ] @xmath106 has no periodic points . suppose @xmath536 is a periodic point . take @xmath481 such that @xmath430 lies on a leaf @xmath292 with @xmath537 . then @xmath538 as @xmath163 and @xmath539 for all @xmath152 , a contradiction . now consider a point @xmath540 . take leaves @xmath541 and @xmath542 such that @xmath543 . by a compactness argument , one can show that @xmath544 has a finite number of connected components , each of which is a properly embedded line tangent to the center bundle of @xmath106 . thus , if @xmath171 is the component containing @xmath146 , there is @xmath27 such that @xmath545 . as @xmath546 is a fixed - point - free homeomorphism of a line , the orbit @xmath547 is unbounded in @xmath171 . using the coordinates @xmath528 for points in @xmath526 , write @xmath548 . unboundedness inside @xmath171 implies that @xmath549 and @xmath550 . expansiveness of the anosov flow can be used to show that there is a unique leaf @xmath374 such that @xmath551 this defines the map @xmath381 . one can verify that the fibers of @xmath383 are exactly the leaves of @xmath262 . hence , @xmath262 is a true foliation and is the unique invariant foliation tangent to the center - stable direction of @xmath106 . a similar map @xmath552 exists and their product gives a homeomorphism between the space of center leaves of @xmath106 and the orbits of the flow @xmath529 . using the averaging method of fuller with @xmath553 , one can construct a leaf conjugacy . we now discuss reebless foliations on seifert spaces and give an outline of the proof of [ thm : phseifert ] . for simplicity , we only consider circle bundles , instead of seifert spaces containing exceptional fibers . we also need only consider circle bundles over surfaces of genus @xmath85 , since circle bundles over @xmath554 and @xmath129 have nilpotent fundamental group . suppose @xmath57 is a circle bundle over a surface @xmath555 . the circle bundle is given by the fibers of a map @xmath556 . call an immersed surface @xmath171 _ horizontal _ if @xmath557 is a local homeomorphism . call @xmath171 _ vertical _ if @xmath558 where @xmath147 is an immersed topological circle or line in @xmath555 . brittenham proved the following result @xcite . [ thm : brittenham ] let @xmath57 be a circle bundle over a surface @xmath555 and let @xmath37 be a reebless foliation on @xmath57 . then , the bundle map @xmath559 may be chosen so that every leaf of @xmath37 is either horizontal or vertical . the regularity of the foliation is crucial . builds on unpublished work of thurston , who showed that if the reebless foliation @xmath37 is @xmath339 , then every leaf is horizontal . see @xcite for details . in @xmath32 regularity or lower , it is possible to have a mixture of horizontal and vertical leaves . see ( * ? ? ? * chapter 4 ) for examples , as well as an introduction to subject and a simplified outline of the proof of [ thm : brittenham ] . in the partially hyperbolic setting , the branching foliation @xmath69 or its approximation @xmath180 is @xmath195 with @xmath32 leaves . however , we may use the dynamics to show that a mix of horizontal and vertical leaves is impossible . [ prop : novert ] let @xmath9 be a transitive partially hyperbolic diffeomorphism defined on a circle bundle @xmath57 over a surface @xmath555 of genus @xmath560 . then every leaf of @xmath69 is horizontal . for simplicity , assume @xmath9 is dynamically coherent , so that @xmath69 is a true foliation . applying [ thm : brittenham ] , we may consider each leaf of @xmath69 as either horizontal or vertical . let @xmath561 be the union of all vertical leaves in @xmath69 . note that @xmath562 , since if every leaf of a foliation on @xmath57 were vertical , it would imply that there is a one - dimensional foliation on the higher genus surface @xmath555 . to show there are no vertical leaves we assume @xmath563 and derive a contradiction . let @xmath239 be small and let @xmath564 denote the unstable segment of length @xmath565 centered at @xmath174 . since @xmath9 maps vertical leaves to vertical leaves , @xmath2 . one can also show that @xmath1 is a closed subset of @xmath57 . if @xmath566 , then there is @xmath239 such that @xmath567 satisfies @xmath568 . then , @xmath569 for all large @xmath137 and this contradicts transitivity . [ idea of the proofs of [ thm : phseifert ] and [ cor : noproduct ] . ] since all of the leaves of @xmath69 are horizontal , they are transverse to the fibers of the seifert fibering . in particular , no unit vector in @xmath71 is tangent to the fibering . assuming that @xmath61 is orientable , this gives us a vector field @xmath570 which is nowhere tangent to the fibering . assume that the bundle map @xmath556 has derivative @xmath571 and so the vector field @xmath572 composed with @xmath573 gives a non - zero map @xmath574 . note that at this point we can prove [ cor : noproduct ] directly . suppose @xmath57 is the direct product @xmath575 . then , there is a section @xmath576 . for instance , map @xmath555 to @xmath577 . composing the map @xmath578 with this section yields a non - vanishing vector field on @xmath555 . since @xmath555 is a higher - genus surface , this is impossible . we now go back to the general case of [ thm : phseifert ] . our goal is to show that @xmath57 supports an anosov flow . to do this , one uses @xmath579 to define a map from @xmath57 to the unit tangent bundle @xmath580 taking fibers to fibers . the properties of circle bundles imply that there is a finite covering map from @xmath57 to @xmath580 . then , the geodesic flow on @xmath580 lifts to an anosov flow on @xmath57 . at the time of writing , it is not clear if the assumption of transitivity is necessary in [ thm : phseifert ] . it seems possible that a careful analysis of the possible dynamics on vertical @xmath581 and @xmath582-leaves could be used to establish [ thm : phseifert ] for all partially hyperbolic systems . after seifert spaces , the next `` natural '' family of manifolds to consider are ones with hyperbolic geometry ( or atoroidal ) . goodman demonstrated that some hyperbolic 3-manifolds support anosov flows @xcite , see also @xcite . there is an established history of research for reebless foliations on hyperbolic 3-manifolds ( * ? ? ? * chapters 7 and 8) . however , the study of partially hyperbolic systems on these manifolds has only barely begun ( * ? ? ? * section 3 ) . for instance , many hyperbolic 3-manifolds , such as the weeks manifold , do not support reebless foliations and so do not support partially hyperbolic diffeomorphisms @xcite . we refer the reader to @xcite and references therein for results concerning anosov and pseudo - anosov flows in 3-manifolds which are not atoroidal . so far , we have only considered restrictions on the manifold @xmath57 . we now consider assumptions on the dynamics itself . we first discuss a very strong condition for the invariant foliations called quasi - isometry . consider a foliation @xmath37 on a manifold @xmath57 . lift @xmath37 to @xmath322 on the universal cover . for two points @xmath583 on the same leaf @xmath470 one can define a distance @xmath584 as the length of the shortest path inside the leaf @xmath171 . ignoring @xmath171 , one can also define @xmath585 as distance on the ambient manifold @xmath258 . the foliation @xmath37 is _ quasi - isometric _ if these two notions of distance are roughly the same ; to be precise , if there is a uniform constant @xmath586 such that @xmath587 for any @xmath588 lying on the same leaf of @xmath322 . the use of the word `` quasi - isometric '' here is slightly different than the much more common notion of two metric spaces being quasi - isometric to each other . the concept of a quasi - isometric foliation was first applied to partially hyperbolic systems by brin , who proved the following @xcite . [ thm : qidc ] suppose @xmath0 is absolutely partially hyperbolic and that each of @xmath169 and @xmath168 is a quasi - isometric foliation . then , @xmath71 , @xmath72 , and @xmath61 are uniquely integrable . if @xmath71 is not uniquely integrable , then there is a path @xmath589 \to { \tilde{m}}$ ] piecewise tangent to @xmath71 with endpoints @xmath590 which lie on the same unstable leaf . since @xmath9 is _ absolutely _ partially hyperbolic , there are constants @xmath591 such that @xmath592 for all unit vectors @xmath593 and @xmath53 . quasi - isometry and the expansion along @xmath63 imply that @xmath594 for all @xmath137 . since @xmath147 is piecewise tangent to @xmath71 , @xmath595 for large @xmath137 , these two estimates give a contradiction . the points @xmath596 and @xmath597 lie on a piecewise @xmath32 curve tangent to @xmath71 , shown here as two curved segments joined in a cusp . they also lie on the same unstable curve , shown here as a vertical line . by quasi - isometry of @xmath169 , the distance between the endpoints grows at least on the order of @xmath598 , but the length of the @xmath71 curve grows only on the order of @xmath599 , a contradiction . ] we have given an outline of this proof not only because it is short and elegant , but because it highlights why it is necessary to require absolute instead of pointwise partial hyperbolicity when using the assumption of quasi - isometry . in some sense , the proof is a global version of the local arguments used in establishing [ teo - strongunstablemanifold ] . brin s motivation in proving [ thm : qidc ] was to answer the question of dynamical coherence on the 3-torus . working with burago and ivanov @xcite , he achieved this goal in the end . [ thm : torusdc ] let @xmath301 be absolutely partially hyperbolic . then , each of @xmath169 and @xmath168 is a quasi - isometric foliation . as a consequence , @xmath9 is dynamically coherent . this theorem was later extended by parwani to all nilmanifolds in dimension 3 @xcite . it was also a major motivation for the classification up to leaf conjugacy of partially hyperbolic systems on @xmath75 . in fact , a version of the classification holds for higher dimensional tori when quasi - isometry is assumed . [ @xcite ] [ thm : hightorus ] suppose that @xmath135 is absolutely partially hyperbolic with @xmath600 and that each of @xmath168 and @xmath169 is a quasi - isometric foliation . then @xmath9 is leaf conjugate to a linear automorphism of @xmath134 . many dynamical properties can be established when assuming quasi - isometry of the foliations @xcite . quasi - isometry is such a strong assumption , however , that it is unlikely to hold for the vast majority of partially hyperbolic systems of interest . brin briefly discusses this in @xcite . several conditions which prevent @xmath168 and @xmath169 from being quasi - isometric are given in @xcite . the quasi - isometric condition and absolute domination in theorem [ thm : hightorus ] can be removed if one assumes that @xmath9 is isotopic to a linear anosov automorphism by a path of partially hyperbolic diffeomorphisms ( see @xcite for precise and more general statements , in particular , in dimension 3 , stronger results hold @xcite ) . in @xcite conditions on the center - stable foliation are given that make the diffeomorphism resemble a linear anosov automorphism of a torus or a nilmanifold . we suspect that the 3-torus and other nilmanifolds are the only 3-dimensional manifolds where the quasi - isometry property holds for partially hyperbolic systems . if one also considers quasi - isometry for the center foliation , then the following holds . [ @xcite ] for an absolutely partially hyperbolic system on a 3-manifold , the stable , center , and unstable foliations exist and are quasi - isometric if and only if @xmath57 is finitely covered by the 3-torus . bonatti and wilkinson @xcite investigated the properties of transitive partially hyperbolic systems in dimension 3 . their original plan was to construct new examples of such systems by gluing together pieces of already known examples . for instance , they hoped to construct a system which resembled a skew product in one part of the manifold and the time-1 map of a flow elsewhere . they soon realized that such surgery techniques were not possible in this setting and proved the following theorems . [ thm : transkew ] let @xmath9 be a partially hyperbolic transitive diffeomorphism on a 3-manifold @xmath57 . assume that @xmath601 is a circle tangent to @xmath61 and that @xmath602 contains at least one other connected component which is a circle . then , @xmath9 is finitely covered by a skew product . [ thm : transflow ] let @xmath9 be a dynamically coherent , transitive diffeomorphism on a 3-manifold . assume @xmath601 is a circle tangent to @xmath61 such that @xmath603 for every center leaf @xmath604 . then @xmath9 is topologically conjugate to the time-1 map of a continuous expansive flow . further consequences also hold . see @xcite for the precise results . it is further conjectured in @xcite that the continuous expansive flow obtained in the above theorem is orbit equivalent to an anosov flow . in our view , showing leaf conjugacy to such an expansive flow would be enough for the purposes of the classification of partially hyperbolic diffeomorphisms . the proofs of the above two theorems rely on the following property : if @xmath9 is transitive and @xmath189 is a periodic compact center leaf , then each of @xmath605 and @xmath606 is dense in @xmath57 . the intersections of @xmath605 and @xmath606 may then be analyzed to establish global properties of the dynamics . say that a partially hyperbolic diffeomorphism has a compact center foliation if there is an invariant foliation tangent to @xmath61 consisting entirely of compact leaves . one step in proving [ thm : transkew ] is the following result , which does not rely on transitivity . [ thm : skew3 ] if @xmath9 is a partially hyperbolic system with compact center foliation on a 3-manifold @xmath57 , then @xmath9 is finitely covered by a skew product . this result was generalized independently and at nearly the same time by d. bohnet @xcite , p. carrasco @xcite , and a. gogolev @xcite . if @xmath9 is a partially hyperbolic system with compact center foliation and @xmath607 , then @xmath9 is finitely covered by a skew product . gogolev also proved the following , which applies to systems in dimensions 4 and 5 . [ thm : skew212 ] if @xmath9 is a partially hyperbolic system with compact center foliation , @xmath600 , @xmath608 , and @xmath609 , then @xmath9 is finitely covered by a skew product . an open question , attributed in @xcite to c. c. pugh , asks if this always holds , regardless of the dimensions of the subbundles . is every partially hyperbolic system with compact center foliation finitely covered by a skew product ? we say that a compact foliation is _ uniformly compact _ if there is a uniform bound on the volume of every leaf . this extra condition is not redundant . for instance , sullivan constructed a foliation of a 5-manifold by circles with no uniform bound on the length @xcite . in the partially hyperbolic setting , however , it is an open question if any compact center foliation is also uniformly compact . using uniform compactness as a hypothesis , d. bohnet showed the following . [ thm : skewu1 ] if @xmath9 is a partially hyperbolic system with uniformly compact center foliation and @xmath133 , then @xmath9 is finitely covered by a skew product . note that there is no restriction here on the dimensions of @xmath62 or @xmath61 . see @xcite for further results on compact center foliations . examples of partially hyperbolic diffeomorphisms ( in any dimension ) appear in four basic forms . 1 . algebraic and geometric constructions , including linear automorphisms of tori and nilmanifolds and geodesic and frame flows ; 2 . skew - products , 3 . surgery on existing examples , 4 . deformation and composition preserving cone - fields . none of the four mechanisms is completely understood . algebraic examples are not even completely understood for the construction of anosov diffeomorphisms . skew products where the base is more hyperbolic than the fiber can be said to be well understood , but when they work the other way around ( the expansion and contraction are seen in the fibers ) , this is just starting to be studied and several exciting examples are starting to appear ( see @xcite ) . this section gives an overview of new examples of partially hyperbolic diffeomorphisms in 3-manifolds appearing in @xcite . they exploit the fourth mechanism , which in a certain sense generalizes the fact that partially hyperbolic diffeomorphisms form an open set in the space of diffeomorphisms . it depends on the possibility of having a good control on the position of the cone - fields around the invariant bundles . this is the mechanism we will present in this section . the main purpose is to construct new examples starting with time one maps of anosov flows but such that they are not isotopic to the identity . [ prop - mecanism ] let @xmath141 be a partially hyperbolic diffeomorphism with splitting @xmath610 . let @xmath611 be a smooth diffeomorphism such that for every @xmath612 one has : @xmath613 then , there exists @xmath425 such that @xmath614 is partially hyperbolic . to do this , consider first a given cone - field @xmath495 for @xmath9 , that is , @xmath615 and vectors in @xmath495 are expanded by @xmath158 . notice that by considering the cone - fields @xmath616 one can assume that @xmath495 is as narrow as one wishes . by compactness and the fact that @xmath617 is transverse to @xmath618 one can assume that @xmath495 is such that @xmath619 is also transverse to @xmath618 . in particular , there exists @xmath425 such that @xmath620 . if @xmath137 is large enough one can also assume that @xmath621 expands vectors in @xmath495 . the existence of a stable cone - field is almost symmetric ( but not exactly since composition with @xmath94 is on the right ) . indeed , by considering the cone - field @xmath622 where @xmath623 is a stable cone - field for @xmath9 one obtains invariance under @xmath624 . if @xmath57 is a 3-manifold whose fundamental group has exponential growth , then it can not admit an anosov diffeomorphism or a skew - product . if @xmath57 admits an anosov flow , then the fundamental group must have exponential growth . therefore , to find new examples of partially hyperbolic diffeomorphisms it is enough to start with an anosov flow @xmath625 of a 3-manifold @xmath57 and find a diffeomorphism @xmath94 which verifies the conditions of [ prop - mecanism ] and such that no iterate of @xmath94 is isotopic to the identity . consider an anosov flow @xmath626 transverse to a torus @xmath627 . one example of such a flow is the suspension of an anosov diffeomorphism of the torus , but wilder examples exist ( see @xcite and references therein ) and it will be those that will interest us . the reason is the following : we plan to compose the time @xmath628-map of @xmath625 with a dehn twist along a neighborhood of @xmath333 so that conditions are satisfied but also so that the dehn twist is not isotopic to the identity ; when considering the suspension of an anosov diffeomorphism of the torus , this is not the case as was proved in @xcite . by dehn twist we mean that we consider an embedded submanifold @xmath629 of @xmath57 diffeomorphic to @xmath229 \times \tt^2 $ ] and an integer vector @xmath630 and consider the diffeomorphism of @xmath57 which is the identity outside @xmath629 and in @xmath629 ( with the coordinates @xmath229 \times \tt^2 $ ] ) is given by @xmath631 where @xmath632 \to [ 0,1]$ ] is a smooth function which equals @xmath633 in a neighborhood of @xmath633 and @xmath282 in a neighborhood of @xmath282 . the easiest example that will work for us was constructed by franks and williams ( @xcite ) and is an anosov flow @xmath625 which is transverse to a torus @xmath333 and has the property that @xmath634 for every @xmath635 . this implies in particular that the flow @xmath625 is not transitive . the dehn twist is constructed so that it is supported on a set of the form @xmath636 for very large @xmath628 so that the effect of the modification can be made to be almost negligible at the level of the derivative , this allows the twist to satisfy conditions . the details can be found in @xcite but let us say a few more words on how this is done : * parametrize @xmath637 by a diffeomorphism @xmath638 \to \mathcal{u}_n$ ] defined by @xmath639 since the center - stable foliation is invariant under the flow , preimages of its leaves under @xmath640 are sets of the form @xmath641 $ ] where @xmath189 is a curve in @xmath333 . therefore , the center - stable foliation on @xmath57 defines a one - dimensional foliation @xmath642 on @xmath333 . similarly , a transverse foliation @xmath643 exists for the center - unstable direction . * if the example is done carefully , it is possible to construct a path of diffeomorphisms @xmath644}$ ] of @xmath333 which is not homotopically trivial and such that @xmath645 is transverse to @xmath643 and @xmath646 is transverse to @xmath642 for all @xmath647 . one requires the path to start and end in the identity and such that for every @xmath648 one has that the curve @xmath649 is homotopically non - trivial . * by choosing a very large @xmath628 , one has that the strong bundles are very close to being tangent to the tangent spaces of the one dimensional foliations @xmath650,@xmath651 in the @xmath652 $ ] coordinates . therefore , in these coordinates , one can construct a dehn twist @xmath653 which , when sent back to @xmath637 via @xmath640 will verify the conditions . by general 3-manifold topology reasons , one can show as desired that this dehn twist will not be homotopic to the identity . in @xcite a specific example is treated where one can show this via elementary methods ( showing that the action in homology is non - trivial ) . recently , in @xcite it is shown that this construction can be made to any non - transitive anosov flow . the construction in the previous subsection relied on the fact that @xmath654 is disjoint from @xmath333 for all @xmath655 . this allowed us to spread out the perturbation and obtain the transversality conditions . if one wishes to make a transitive example starting from a transitive anosov flow , this condition will never be satisfied . one of the main constructions in @xcite consists in making a quantitative version of the previous argument , applied to certain transitive anosov flows transverse to tori ( see @xcite for plenty of such examples ) that after huge finite lifts will be treated in a similar way as the non - transitive case . this implies solving several technical difficulties , but at the end it is possible to do so without so many differences as in the previous case . to guarantee that after perturbation the new diffeomorphism is transitive , one could argue directly with some attention to be able to use the transitivity of the anosov flow and the control on the perturbation , but there is a way to do this with an easy trick that moreover allows us to obtain even better results ( construct stably ergodic examples ) . the key point is that if the anosov flow is volume preserving and the dehn twist with which we compose preserves the same volume form , the new example will be also conservative , and therefore easier to show it can be perturbed to be robustly transitive ( and stably ergodic ) . to do this , in ( * ? ? * section 3 ) the example is constructed starting with the anosov flow constructed in @xcite and as explained before , considering a big finite cover . in @xcite there are examples of different nature also , constructed via dehn twists along tori , but which are not transverse to the anosov flow . the transversality conditions can be guaranteed by `` geometric '' reasons . we will explain these examples in the following subsection . let @xmath657 be any smooth diffeomorphism . the projectivization @xmath658 defined as @xmath659 conjugates the geodesic flow of the metric @xmath14 with the one obtained by the pullback @xmath660 . that is , if @xmath661 denotes the geodesic flow of @xmath660 one has that : [ prop - closemetric ] it is possible to construct @xmath657 not isotopic to the identity in @xmath210 so that @xmath664 and @xmath14 are very close to each other ( in the @xmath665-topology ) and with @xmath14 having constant negative curvature . this is done in ( * ? ? ? * section 2 ) via considering a hyperbolic metric of constant curvature @xmath666 with a very short closed geodesic and choosing @xmath9 to be a dehn twist along this curve , since the geodesic is short , it follows that , in the universal cover , the lift of the diffeomorphism @xmath9 is very close to the identity . it follows that the diffeomorphism @xmath667 for large @xmath628 will be partially hyperbolic thanks to proposition [ prop - mecanism ] and not isotopic to the identity because @xmath663 is not . since @xmath663 does not preserve the length along fibers , it will not preserve the volume form on @xmath668 . however , it is possible to find a diffeomorphism @xmath32-close to @xmath663 which does preserve this volume form and for which [ prop - mecanism ] still applies . in principle , this idea and the examples presented in the previous section suggest that one might be able only to make examples isotopic to dehn twists . these have the particularity that their action in homology only has eigenvalues of modulus @xmath282 . the proof of proposition [ prop - closemetric ] as explained above , requires a small geodesic along which one performs a dehn twist . if one would like to construct @xmath9 which belongs to a mapping class having eigenvalues larger than @xmath282 , then one would need to compose with crossing dehn twists , but if one curve is short , the transverse one should be long . in @xcite this problem is solved by considering maps and geodesic flows in non - connected surfaces and choosing the initial diffeomorphisms not to be the geodesic flows , but to permute the connected components and apply the geodesic flow so that the diffeomorphism visits every connected component . in this way , it is possible to modify the metrics little by little in order to apply dehn twists in different connected components so that a sufficiently large iterate will observe the multiple dehn twists and the return map to a connected component will be partially hyperbolic . in particular , it is shown : [ teo - unittangentbundle ] given a surface @xmath210 of genus @xmath669 and a diffeomorphism @xmath657 , there exists a ( stably ergodic ) partially hyperbolic diffeomorphism @xmath670 isotopic to @xmath663 . the mechanism for constructing new examples needs to start with a partially hyperbolic diffeomorphism on @xmath57 . in particular , this can only give examples on new isotopy classes , but not on different manifolds . in this direction recall [ thm : transflow ] stating that if @xmath9 is transitive and every center - leaf is fixed by @xmath9 and there is a circle leaf then this must be essentially the case ( see the discussion after the statement of [ thm : transflow ] ) . see @xcite for more questions arising naturally from these new examples . c. bonatti , l. daz and m. viana , _ dynamics beyond uniform hyperbolicity . a global geometric and probabilistic perspective _ , encyclopaedia of mathematical sciences * 102*. mathematical physics iii . springer - verlag ( 2005 ) . m. brin , d. burago and s. ivanov , on partially hyperbolic diffeomorphisms of 3-manifolds with commutative fundamental group . _ modern dynamical systems and applications _ cambridge univ . press , cambridge , ( 2004 ) 307312 . m. gromov , three remarks on geodesic dynamics and fundamental group , preprint suny ( 1976 ) , reprinted in _ lenseignement mathmatique _ * 46 * ( 2000 ) 391402 . m. gromov , groups of polynomial growth and expanding maps . ihes _ , * 53 * ( 1981 ) 5373 . a. hammerlindl , leaf conjugacies in the torus , _ ergodic theory and dynamical systems _ * 33 * 3 ( 2013 ) 896 933 . m.a , rodriguez hertz , f. rodriguez hertz and r. ures , a survey on partially hyperbolic systems , _ fields institute communications _ , partially hyperbolic dynamics , laminations and teichmller flow , * 51 * ( 2007 ) 3588 .

this paper surveys recent results on classifying partially hyperbolic diffeomorphisms . this includes the construction of branching foliations and leaf conjugacies on three - dimensional manifolds with solvable fundamental group . classification results in higher - dimensional settings are also discussed . the paper concludes with an overview of the construction of new partially hyperbolic examples derived from anosov flows . * keywords : * partial hyperbolicity , dynamical coherence , leaf conjugacy . * msc 2010 : * 37c05 , 37d30 , 57r30 .

it has been known for a long time that the energy density of a quantum field may assume negative values @xcite . if the magnitude and the duration of such negative energy densities were unconstrained various exotic phenomema might be possible , ranging from the manufacture of macroscopic traversable wormholes @xcite to large - scale violations of the second law of thermodynamics @xcite . however , quantum field theory itself contains mechanisms to limit the extent of negative energy densities : certain weighted averages of the stress - energy tensor obey state - independent lower bounds known as quantum weak energy inequalities ( qwei ) [ or simply quantum inequalities ( qi ) ] which severely limit exotic phenomena such as those described above . bounds of this type have been obtained by various means @xcite for the scalar , electromagnetic and proca fields , in both flat and curved spacetimes , leading to results of great generality . taking four - dimensional minkowski space as a specific example , and averaging the energy density along an inertial worldline , the qwei satisfied by these three theories may be written in the form @xcite @xmath1 for any smooth , real - valued @xmath2 vanishing outside a compact region , where @xmath3 is the particle s mass , @xmath4 denotes its helicity ( @xmath5 for scalars , @xmath6 for photons and @xmath7 for massive spin-@xmath8 particles ) , while @xmath9 and is replaced by @xmath10 in the massless case . ( in curved spacetimes the qweis of these theories are not simply related by an overall factor . ) in the spin-@xmath11 case only two arguments are known . one , due to vollick @xcite ( based on earlier work of flanagan @xcite ) , exploits the conformal properties of massless dirac fields in two - dimensional spacetimes to obtain a sharp ( and explicit ) qwei for an arbitrary sampling function . the other , due to fewster and verch @xcite , establishes the existence of qewis for dirac and majorana fields of arbitrary mass in general four - dimensional globally hyperbolic spacetimes by drawing on techniques from microlocal analysis . however , this approach did not lead to an explicit formula for the qwei bound , so its magnitude even in four - dimensional minkowski space has not previously been determined . ( specific classes of dirac states with local negative energy densities have been studied @xcite and shown , in the massless limit , to satisfy the scalar field qwei obtained in @xcite ; of course , this does not amount to a proof of the qweis . ) thus , for example , it has not previously been clear whether or not it is more or less easy to support a traversable wormhole using dirac fields , rather than with the scalar or electromagnetic fields . in this paper , we present a new qwei derivation for dirac fields , which combines ideas taken from @xcite with an approach first developed for the scalar field in @xcite . the argument is straightforward , involving the use of anticommutation relations and an identity for fourier transforms , and yields an explicit , closed - form , bound . in four - dimensional minkowski space this bound is @xmath12 where @xmath13 is replaced by @xmath10 in the massless case . although we have restricted ourselves , for simplicity , to minkowski space , our technique certainly extends to static spacetimes and may even suggest a new approach in the general globally hyperbolic case . it is hoped to return to these issues elsewhere . + _ conventions : _ we take the metric to have signature @xmath14 and adopt units in which @xmath15 . the fourier transform @xmath16 of a function @xmath17 on @xmath18 is defined by @xmath19 to establish our notation , we begin by recalling some basic facts about the dirac fields . the dirac equation for a fermion field @xmath20 of mass @xmath3 is @xmath21 where the @xmath22-matrices are given in terms of the pauli matrices @xmath23 by @xmath24 and obey @xmath25 . the stress - energy tensor @xcite is @xmath26,\ ] ] where @xmath27 , so the classical dirac energy density is @xmath28 . \label{eq : t00}\ ] ] ( of course , this quantity is unbounded both from above and below , in contrast to the energy density of a classical scalar field . this is at the root of the differences between the qweis obeyed by the scalar and dirac fields . ) to simplify our notation , we quantise the dirac field in a box of side @xmath29 and then take the continuum limit @xmath30 towards the end of the argument . we emphasise that this is done purely for notational convenience . the dirac field operator @xmath31 is therefore defined in terms of creation and annihilation operators by @xmath32 , \label{eq : psi}\ ] ] where @xmath33 runs through @xmath34 , and the four - vector @xmath35 has components @xmath36 with @xmath37 . the annihilation and creation operators satisfy the anticommutation relations @xmath38 and @xmath39 with all other anticommutators vanishing . finally , the spinors @xmath40 and @xmath41 , where @xmath42 labels the two independent spin states , are given by @xmath43 \displaystyle \frac{{{\boldsymbol{\sigma}}}\cdot { { \boldsymbol{k}}}}{\sqrt{2 \omega_{{\boldsymbol{k}}}(\omega_{{\boldsymbol{k}}}+ m ) v } } \phi^{\alpha } \end{pmatrix } \quad\hbox{and}\quad v^{\alpha}({{\boldsymbol{k}}})= \begin{pmatrix } \displaystyle \frac{{{\boldsymbol{\sigma}}}\cdot { { \boldsymbol{k}}}}{\sqrt{2 \omega_{{\boldsymbol{k}}}(\omega_{{\boldsymbol{k}}}+ m ) v } } \phi^{\alpha } \\[0.4 cm ] \displaystyle \sqrt{\frac{\omega_{{\boldsymbol{k}}}+m}{2 \omega_{{\boldsymbol{k}}}v } } \phi^{\alpha } \end{pmatrix},\ ] ] in which the two dimensional column vectors @xmath44 are @xmath45 and @xmath46 and @xmath47 . the normalisation has been chosen so that @xmath48 substituting the quantum field @xmath31 into the classical energy density eq . and normal ordering according to the prescription @xmath49 we obtain the normal ordered energy density at the spatial origin @xmath50 in the form @xmath51 \nonumber \\ & & \indent + ( \omega_{{{\boldsymbol{k'}}}}-\omega_{{{\boldsymbol{k } } } } ) [ d_{\alpha } ( { { \boldsymbol{k } } } ) b_{\alpha ' } ( { { \boldsymbol{k ' } } } ) v^{\alpha \dagger}({{\boldsymbol{k } } } ) u^{\alpha'}({{\boldsymbol{k'}}})e^{-i(\omega_{{{\boldsymbol{k}}}}+\omega_{{{\boldsymbol{k'}}}})t } \nonumber \\ & & \indent- b_\alpha^\dagger ( { { \boldsymbol{k } } } ) d_{\alpha'}^{\dagger}({{\boldsymbol{k ' } } } ) u^{\alpha \dagger}({{\boldsymbol{k } } } ) v^{\alpha'}({{\boldsymbol{k'}}})e^{i(\omega_{{{\boldsymbol{k}}}}+\omega_{{{\boldsymbol{k'}}}})t } ] \}. \end{aligned}\ ] ] we will consider weighted energy density averages of the form @xmath52 measured by a stationary observer at spacetime origin @xmath53 , where @xmath17 is a non - negative sampling function . thus , @xmath54 \nonumber \\ & & \indent + ( \omega_{{{\boldsymbol{k'}}}}-\omega_{{{\boldsymbol{k } } } } ) [ d_{\alpha } ( { { \boldsymbol{k } } } ) b_{\alpha ' } ( { { \boldsymbol{k ' } } } ) v^{\alpha \dagger}({{\boldsymbol{k } } } ) u^{\alpha'}({{\boldsymbol{k ' } } } ) \widehat{f}(\omega_{{{\boldsymbol{k}}}}+\omega_{{{\boldsymbol{k ' } } } } ) \nonumber\\ & & \indent- b_\alpha^\dagger ( { { \boldsymbol{k } } } ) d_{\alpha'}^{\dagger}({{\boldsymbol{k ' } } } ) u^{\alpha \dagger}({{\boldsymbol{k } } } ) v^{\alpha'}({{\boldsymbol{k ' } } } ) \widehat{f } ( -\omega_{{{\boldsymbol{k}}}}-\omega_{{{\boldsymbol{k ' } } } } ) ] \}. \label{eq : tf } \end{aligned}\ ] ] our aim is to determine a lower bound on expectation values of @xmath55 in the case where @xmath56 , for some real - valued , smooth , compactly supported function @xmath2 . to this end , we first define a family @xmath57 of fock space operators by @xmath58 a short calculation , using the anticommutation relations and the normalisation , shows that @xmath59 where @xmath60 and we have implicitly summed over the spinor index @xmath61 . to relate this to our expression for @xmath55 we use the following lemma , which plays an analogous role here to that played by the convolution theorem in @xcite . + * lemma 1 : * _ if @xmath56 for some real - valued , smooth , compactly supported @xmath62 then _ @xmath63 this lemma was also used in the qwei derivation in @xcite ; for completeness we give a proof of this statement in the appendix which differs slightly from that given in that reference . using this lemma with the anticommutation relations eq . and eq . , we calculate @xmath64 \notag \\ & & \phantom{\sum\sum } + ( \omega_{{{\boldsymbol{k'}}}}-\omega_{{{\boldsymbol{k } } } } ) [ d_{\alpha } ( { { \boldsymbol{k } } } ) b_{\alpha'}({{\boldsymbol{k'}}})v^{\alpha \dagger}({{\boldsymbol{k}}})u^{\alpha'}({{\boldsymbol{k ' } } } ) \widehat{f}(\omega_{{{\boldsymbol{k}}}}+\omega_{{{\boldsymbol{k ' } } } } ) \notag \\ & & \phantom{\sum\sum } - b^{\dagger}_{\alpha } ( { { \boldsymbol{k } } } ) d^{\dagger}_{\alpha'}({{\boldsymbol{k'}}})u^{\alpha \dagger}({{\boldsymbol{k}}})v^{\alpha'}({{\boldsymbol{k ' } } } ) \widehat{f}(-\omega_{{{\boldsymbol{k}}}}-\omega_{{{\boldsymbol{k ' } } } } ) ] \ } .\end{aligned}\ ] ] comparing with eq . , we conclude that @xmath65 which is similar in form to the decomposition of the scalar field energy density employed in @xcite . next we compute the anticommutator @xmath66 where we have again implicitly summed over the spinor index @xmath61 and used the fact that anticommutators between the @xmath67 and @xmath68 vanish . using the anticommutation relations , the above expression reduces to @xmath69 where we have also used eq . and the fact that @xmath70 is even ( because @xmath2 is real ) . splitting the @xmath71-integral of eq . into two pieces and using eq . gives @xmath72 \notag \\ & = \frac{1}{2\pi } \int_{0}^{\infty } d\mu \ , \mu ( \mathcal{o}_{\mu i}^{\dagger } \mathcal{o}_{\mu i } - s_{\mu}{\openone } ) + \frac{1}{2\pi } \int_{-\infty}^{0 } d\mu \ , \mu ( s_{-\mu}{\openone}- \mathcal{o}_{\mu i } \mathcal{o}_{\mu i}^{\dagger } ) . \end{aligned}\ ] ] now , for any @xmath73 , we have @xmath74 , while for any @xmath75 , we have @xmath76 for all physically reasonable states @xmath20 . thus , @xmath77 so we have obtained a state - independent lower bound on @xmath78 . we now proceed to calculate the right - hand side of eq . and show that it is finite . inserting the definition eq . of @xmath79 into eq . , we find @xmath80 taking the continuum limit @xmath81 the inequality becomes @xmath82 where we have made a change of variables in the integration to spherical polar coordinates , integrated over the angular variables and changed variable again to the energy @xmath83 . the next step is to make a further change of variable @xmath84 after which the quantum inequality eq . becomes @xmath85 where the function @xmath86 is given in eq . . using the translational invariance of the theory to move from a worldline @xmath50 to @xmath87 , we have therefore established the bound @xmath88 for any @xmath89 , where @xmath90 as claimed in the introduction . this bound is clearly finite for any smooth compactly supported @xmath2 , as @xmath91 grows like @xmath92 for large @xmath93 , while @xmath94 decays faster than any inverse polynomial in @xmath93 as @xmath95 . we now look at the simplest case of the general qwei derived above ; the massless case . we then compare our result for the dirac field with those for the scalar field derived by ford and roman @xcite and fewster and eveson @xcite . substituting @xmath96 in the first line of eq . and performing the integral over @xmath97 , we immediately find that the four - dimensional qwei for massless fields takes the form @xmath98 which is also the @xmath99 limit of eq . because @xmath100 as @xmath101 . since the integrand is an even function in @xmath93 , we may extend the range of integration to the whole of @xmath18 to give @xmath102 where we have used parseval s theorem to yield a @xmath103-space version of the quantum inequality in the last line . in passing , we observe that , because @xmath104 is bounded between @xmath105 and @xmath8 ( see fig . [ fig : q3d ] ) , the introduction of a mass only serves to tighten the qwei bound . accordingly , eq . also holds for massive dirac fields . comparison with the quantum inequality for the scalar and electromagnetic fields in minkowski space derived in @xcite shows that the massless dirac bound is weaker than the corresponding scalar bound by a factor of @xmath106 , and stronger than the electromagnetic bound by a factor of @xmath107 . since none of these bounds are optimal , it is difficult to draw definite conclusions from this beyond the observation that the bounds are of comparable magnitude . in particular , for the choice of a lorentzian sampling function peaked at @xmath108 , with characteristic width @xmath109 @xmath110 in which @xmath109 sets the averaging timescale , it is straightforward to calculate @xmath111 ( at least for those states for which the right - hand side is well - defined ) . this is a slightly stronger result , by a factor of @xmath112 , than the quantum inequality derived by ford and roman @xcite for the scalar field . we have therefore explained why the specific classes of states considered in @xcite satisfy the ford roman qwei . we also observe that a traversable wormhole supported by a dirac field would be subject to slightly tighter constraints than those envisaged in @xcite . we have derived a new and explicit qwei for the dirac field of mass @xmath0 in four - dimensional minkowski space , by straightforward means . our argument is related to the general ( but inexplicit ) qwei obtained in @xcite , in that lemma 1 is at its heart ; one may therefore regard it as an improved and streamlined version of @xcite ( for the particular case of averaging along an inertial trajectory in minkowski space ) with the added benefit of a closed - form expression for the bound . although it is at present somewhat formal , we expect that our argument can be made fully rigorous ( by means of microlocal techniques , cf . @xcite ) and that it can shed further light on the general situation in curved spacetimes , again leading to explicit bounds . our argument is also related to that developed for the scalar field in @xcite , and which we briefly summarise in appendix [ appx : b ] . in each case , an identity for fourier transforms ( the convolution theorem or lemma 1 ) and the ( anti)commutation relations are used to express the averaged energy density as an integral of manifestly positive operators , modulo a @xmath113-number term which eventually provides the lower bound . extra care is required in the dirac case because the classical energy density is unbounded both from above and below . finally , in contrast to vollick s two - dimensional bound @xcite , our qwei is not expected to be the sharpest bound possible . again , it is hoped to address this issue elsewhere . in this appendix , we wish to state and prove the following lemma ( originally proved in @xcite by a slightly different method ) : if @xmath56 for real - valued @xmath62 , then @xmath114 to obtain this result , first notice that the left - hand side is the momentum space integral kernel for the operator @xmath115 where @xmath17 acts by multiplication and @xmath116 . that is , @xmath117 holds ( at least ) for all smooth , compactly supported @xmath20 . on the other hand , because @xmath56 , we also have @xmath118g+g[g , p ] = 2gpg\ ] ] because @xmath119=ig'$ ] , another multiplication operator . written as an integral operator , we therefore have @xmath120 since both and hold for all smooth compactly supported @xmath20 , the two integral kernels must also agree , therefore yielding , using the fact that @xmath121 because @xmath2 is real . for ease of comparison with our dirac qwei , we briefly summarise the scalar field argument of @xcite . ( notation differs slightly , and we have specialised to four - dimensional minkowski space . ) this argument begins by observing that the averaged energy density may be written as a sum of terms of the form @xmath122 where @xmath123 and @xmath124 are the annihilation and creation operators of the scalar field , and @xmath125 is polynomially bounded in @xmath35 . introducing operators @xmath126 a calculation using the commutation relations satisfied by the annihilation and creation operators and the convolution theorem in the form @xmath127 ( and the fact that @xmath128 as @xmath2 is real ) yields the expression @xmath129 since the right - hand side of eq . is ( formally ) a manifestly positive operator , we conclude that the expectation value of @xmath130 in any quantum state @xmath20 must satisfy the following bound : latexmath:[\[\langle s^\pm\rangle_{\psi } \geqslant - \int_0^\infty \frac{d\omega}{2\pi } \int \frac{d^{3}{{\boldsymbol{k}}}}{(2\pi)^{3 } }

quantum weak energy inequalities ( qweis ) have been established for a variety of quantum field theories in both flat and curved spacetimes . dirac fields are known ( by a result of fewster and verch ) to satisfy qweis under very general circumstances . however this result does not provide an explicit formula for the qwei bound , so its magnitude has not previously been determined . in this paper we present a new and explicit qwei bound for dirac fields of mass @xmath0 in four - dimensional minkowski space . we follow the methods employed by fewster and eveson for the scalar field , modified to take account of anticommutation relations . a key ingredient is an identity for fourier transforms established by fewster and verch . we also compare our qwei with those previously obtained for scalar and spin-1 fields .

recently , research on the spinel compound liv@xmath0o@xmath1 has been an exciting field since the discovery of the large electronic specific heat coefficient @xmath3 = 420 mj / mol@xmath4k@xmath5 at low temperature @xcite . this value of @xmath3 is the largest among @xmath6-electron compounds . moreover , liv@xmath0o@xmath1 has been found to exhibit heavy - fermion like behavior ; i.e. , a broad maximum in the magnetic susceptibility @xmath7 @xcite , @xmath8-dependence in the electrical resistivity @xcite , and the korringa type relaxation , @xmath9 , where @xmath10 is the spin - lattice relaxation rate derived from the @xmath11li - nmr measurements @xcite . it should also be noted that the solid - solution li@xmath12zn@xmath13v@xmath0o@xmath1 undergoes a metal - insulator transition as a function of @xmath14 @xcite . in spite of enormous studies , however , origin of these anomalous behavior is not clearly solved . several models have been proposed to explain the physical properties of liv@xmath0o@xmath1 . at first , the large @xmath3 has been attributed to the kondo - lattice formation as in the case of @xmath15-electron based compounds @xcite . have shown that their @xmath11li - nmr results are consistent with this mechanism . however , temperature dependence of the electrical resistivity at high temperatures differs qualitatively from that of the @xmath15-electron based kondo - lattice systems @xcite . for the kondo - lattice formation in liv@xmath0o@xmath1 , both the conduction and the localized electrons are needed , whereas the presence of well - defined localized electrons is not clearly demonstrated . on the other hand , fujiwara _ et al . _ have interpreted their @xmath11li - nmr results as the characteristic behavior of itinerant - electron systems with strongly - enhanced ferromagnetic spin - fluctuations @xcite . they have shown that the temperature dependence of @xmath10 is well described by the spin - fluctuation theory ( scr theory ) for weakly- or nearly - ferromagnetic metals @xcite . neutron scattering studies by krimmel _ have also suggested possible nearly - ferromagnetic state for liv@xmath0o@xmath1 at high temperatures @xcite . thus , physical properties of liv@xmath0o@xmath1 have been argued in terms of completely different electronic states , i.e. , localized- or itinerant - electron systems . in addition , the effect of geometrical frustration arising from the spinel structure is suggested to be essential for the anomalous behavior in liv@xmath0o@xmath1 @xcite . this is inferred from the observations of spin - glass order in a slightly mg- or zn - doped variants @xcite . to obtain further insight for the magnetic state of liv@xmath0o@xmath1 , we have performed a high - field magnetization measurement on liv@xmath0o@xmath1 . since magnetic field can enhance the ferromagnetic spin - correlation , this experiment would give valuable information about the spin correlation in liv@xmath0o@xmath1 . to see the effect of geometrical frustration , we have also studied on an oxygen - deficient sample liv@xmath0o@xmath2 . powder sample of liv@xmath0o@xmath1 was prepared by a solid state reaction from powder samples of li@xmath16vo@xmath1 , v@xmath0o@xmath16 , and v@xmath0o@xmath17 . the same procedure reported in ref . @xcite was employed . an oxygen deficient sample liv@xmath0o@xmath2 was prepared by varying the nominal composition . x - ray diffraction confirmed the spinel structure for both the samples . magnetic susceptibility was measured using a superconducting quantum interference device ( squid ) magnetometer at 1 koe . high - field magnetization measurements were performed at 4.2 k by an induction method with well - balanced pickup coils . magnetic fields up to 45 t were generated by a long - pulse magnet at the institute for solid state physics , university of tokyo . figure 1 shows inverse magnetic susceptibility 1/@xmath7 of liv@xmath0o@xmath1 and liv@xmath0o@xmath2 as functions of temperature @xmath18 . for both the systems , @xmath7 follows a curie - weiss behavior at high temperatures , and shows no evidence of magnetic ordering down to 2 k. 1/@xmath7 of liv@xmath0o@xmath1 shows a saturation behavior around 10 k , and slightly decreases below 5 k. this small decrease is attributed to the contribution of impurities and/or defects . subtracting this contribution by a curie term yields a minimum at @xmath19 = 16 k , as can be seen in the dotted line in fig . this is consistent with those reported by several authors @xcite . 1/@xmath7 of liv@xmath0o@xmath2 also shows a curie - weiss behavior , and the data above 60 k agree with those of liv@xmath0o@xmath1 . below about 20 k , 1/@xmath7 of liv@xmath0o@xmath2 decreases rapidly without any anomaly . absence of anomaly in @xmath7 of liv@xmath0o@xmath2 has been suggested previously @xcite . one may consider that a large contribution from defects masks a minimum in the intrinsic @xmath20 of liv@xmath0o@xmath2 . however , we found that a minimum is not given in @xmath20 of liv@xmath0o@xmath2 by reasonable corrections . this suggests that @xmath7 of liv@xmath0o@xmath2 at low temperature is qualitatively different from that of liv@xmath0o@xmath1 , though the high - temperature behavior agrees each other . figure 2 shows magnetization @xmath21 of liv@xmath0o@xmath1 and liv@xmath0o@xmath2 as functions of magnetic field @xmath22 . for both the systems , @xmath21 increases linearly up to 15 t. at higher fields , @xmath21 of liv@xmath0o@xmath1 turns to increase nonlinearly , and exhibits an inflection point at @xmath23 = 38 t. such a metamagnetic behavior in liv@xmath0o@xmath1 is also reported in the literature @xcite . magnetic moment at 45 t is about 1.1 @xmath24 per formula unit ( f.u . ) . in contrast , @xmath21 of liv@xmath0o@xmath2 increases monotonically up to 40 t with no indication of metamagnetic behavior . the value of @xmath21 at 40 t is less than 0.7 @xmath24/f.u . , considerably smaller than that of liv@xmath0o@xmath1 . figure 3 shows the arrott - plot of the two systems . the slope of @xmath25 vs. @xmath26 for liv@xmath0o@xmath1 is negative , consistent with the metamagnetic behavior . on the other hand , the slope for liv@xmath0o@xmath2 is positive , in a clear contrast to that for liv@xmath0o@xmath1 . one can therefore expect that @xmath21 of liv@xmath0o@xmath2 increases monotonically until its saturation value , without metamagnetic behavior even for much higher fields . in this section , we argue on the possible mechanism for the metamagnetic behavior in liv@xmath0o@xmath1 . metamagnetic behavior is usually explained as a spin - flopping process in antiferromagnetic ( afm ) ordered systems . in contrast , liv@xmath0o@xmath1 does not order magnetically down to the lowest temperature , as is demonstrated by many works including specific heat @xcite , @xmath11li - nmr @xcite , muon spin relaxation @xcite and neutron diffraction @xcite . hence , the metamagnetic behavior observed in liv@xmath0o@xmath1 is not attributed to a spin - flopping . alternatively , metamagnetic behavior out of a paramagnetic state has been observed in such cases as : ( i ) kondo - lattice systems like ceru@xmath0si@xmath0 @xcite , cecu@xmath27 @xcite or ybcual @xcite , and ( ii ) itinerant - electron systems close to ferromagnetic ( fm ) order @xcite like yco@xmath0 @xcite or tibe@xmath0 @xcite . notably , both the two models , kondo - lattice and nearly - fm metal , have been used to explain the anomalous physical properties of liv@xmath0o@xmath1 . for kondo - lattice systems , the precise mechanism of the metamagnetic behavior has not been clearly understood yet . nevertheless , it is widely accepted that the metamagnetic behavior accompanies a qualitative change in the @xmath15-electron state . for ceru@xmath0si@xmath0 , where a distinct metamagnetic behavior has been observed at @xmath28 = 7 t @xcite , a drastic change in the magnetic response has been demonstrated by the neutron scattering studies @xcite ; from strong afm fluctuations at @xmath29 to single - site ( @xmath30-independent ) fluctuations at @xmath31 . similar change in the magnetic response has also been observed in cecu@xmath27 @xcite . furthermore , the de haas - van alphen effect measurements on ceru@xmath0si@xmath0 have revealed the change in its fermi surface from itinerant at low fields to almost localized @xmath15-electrons at high fields @xcite . hence , the high - field states in these compounds are well described by localized @xmath15-electrons with single - site kondo fluctuations . it is also notable that the magnetization of ce@xmath32(la , y)@xmath13ru@xmath0si@xmath0 at high fields is almost independent of @xmath14 , though the metamagnetic behavior becomes broadened with increasing @xmath14 @xcite . similar behavior is also observed in ce@xmath32la@xmath13fe@xmath0ge@xmath0 @xcite and yb@xmath32y@xmath13cual @xcite . this substitution dependence is well explained by the single - site character of @xmath15-electrons at high fields , since the single - site interactions are insensitive to disorder . thus , we can conclude that such a localized - electron state at high fields is a common feature in kondo - lattice systems . here , note the qualitative difference in the magnetization of liv@xmath0o@xmath1 and liv@xmath0o@xmath2 : @xmath21 of liv@xmath0o@xmath2 at 40 t is considerably smaller than that of liv@xmath0o@xmath1 . this distinct difference in @xmath21 does not agree with the high - field state expected for kondo - lattice systems . therefore , the kondo - lattice model is unlikely for liv@xmath0o@xmath1 as far as the mechanism of the metamagnetic behavior is concerned . instead , this result is likely to suggest that the intersite correlation is still important even at high fields , because intersite correlation should be sensitive to disorder or defects . this indicates that the metamagnetic behavior is relevant to the evolution of intersite fm correlation . thus , we turn to the next picture ; itinerant - electron systems close to fm order . within the itinerant - electron description , the curie - weiss susceptibility in liv@xmath0o@xmath1 is attributed to the temperature dependent fm ( @xmath33 ) spin fluctuations , not to localized moments @xcite . here , curie constants do not necessarily agree with the effective magnetic moment expected for a free ion , and a negative weiss temperature no longer indicates antiferromagnetic ( afm ) interactions . several nearly - fm metals exhibit metamagnetic behavior as well as a maximum in @xmath7 @xcite . such examples are yco@xmath0 @xcite , tibe@xmath0 @xcite , and co(s , se)@xmath0 @xcite . recently , metamagnetic behavior has also been reported in the metallic oxide sr@xmath16ru@xmath0o@xmath34 around @xmath22 = 5 t @xcite . it is notable that this compound shows a maximum in @xmath7 and a large @xmath35 100 mj / ru - mol@xmath4k@xmath5 , and is considered to be close to a fm ordering @xcite . these features of nearly - fm metals are quite similar to those observed in liv@xmath0o@xmath1 including the metamagnetic behavior . on the other hand , it has been shown that liv@xmath0o@xmath1 also exhibits quite contrasting behavior to the feature for nearly - fm metals . for example , neutron scattering experiments by lee _ el al . _ have revealed the presence of afm fluctuations at @xmath36 0.6 @xmath37 @xcite . in addition , spin - glass ordering has been observed in mg- or zn - doped samples @xcite . these observations suggest the presence of strong afm interaction and the effect of the geometrical frustration , the latter of which prevents this system from a long - range afm ordering . in this respect , liv@xmath0o@xmath1 has been compared with ( y , sc)mn@xmath0 and @xmath38-mn @xcite , where the afm order is suppressed by the geometrical frustration , and the resulting fluctuations enhance the electronic specific heat coefficients @xcite . it should , however , be stressed that the temperature dependence of @xmath7 of ( y , sc)mn@xmath0 and @xmath38-mn differs qualitatively from that of liv@xmath0o@xmath1 . @xmath7 of the formers are almost @xmath18-independent and do not show curie - weiss behavior @xcite , consistent with the dominant afm ( @xmath39 0 ) fluctuations . in contrast , @xmath7 as well as the @xmath11li - knight shift @xmath40 of liv@xmath0o@xmath1 show curie weiss behavior , suggesting the importance of the fm ( @xmath41 0 ) fluctuation . furthermore , the metamagnetic behavior observed in the present study evidences the importance of the fm interaction in this system . from these facts , we conclude that both the fm ( @xmath41 0 ) and the afm ( @xmath36 0.6 @xmath37 ) interactions are important in liv@xmath0o@xmath1 . this is similar to the case in v@xmath17se@xmath42 , where both fm ( @xmath41 0 ) and afm ( @xmath43 ) fluctuations are considered to be enhanced @xcite . this model has successfully explained the itinerant afm behavior and the curie - weiss susceptibility of v@xmath17se@xmath42 @xcite . the presence of two kind of spin - fluctuations in liv@xmath0o@xmath1 has been suggested by the earlier neutron scattering study @xcite and the scr analysis for the specific heat data @xcite . recently , eyert _ et al . _ have performed first - principles calculations , and have shown that afm state is stable for liv@xmath0o@xmath1 and the energy of fm state is slightly higher @xcite . this result is indicative that applying high fields would reduce the energy for the fm state , which results in a metamagnetic transition . it is notable that the temperature where @xmath7 has a maximum is @xmath44 10 k for tibe@xmath0 @xcite and @xmath45 = 16 k for sr@xmath16ru@xmath0o@xmath34 @xcite , comparable to @xmath45 = 16 k for liv@xmath0o@xmath1 , whereas the field where metamagnetic behavior occurs is @xmath23 = 6 t for tibe@xmath0 @xcite and @xmath23 = 5 t for sr@xmath16ru@xmath0o@xmath34 @xcite , remarkably smaller than @xmath46 38 t for liv@xmath0o@xmath1 . this may be a result of competition of the fm and the afm fluctuations in liv@xmath0o@xmath1 below @xmath23 . we expect that the afm fluctuations are suppressed at around @xmath23 and the fm correlation becomes dominant above @xmath23 . on the other hand , fm correlation does not appear to develop in liv@xmath0o@xmath2 . this suggests that the afm interaction is dominant even at high fields in this system . instead of the itinerant - electron picture mentioned above , the possibility of liv@xmath0o@xmath1 as a localized - moment system is not still ruled out . the value of the spin - lattice relaxation rate 1/@xmath47 of the @xmath48v @xcite and of the @xmath11li nuclei @xcite at high temperatures are consistent with those for a localized - moment system . the evolution of localized electrons has been suggested theoretically by the splitting of the @xmath49 orbital into almost localized @xmath50 and conductive @xmath51 orbitals due to a trigonally distortion in the vo@xmath27 octahedra and the coulomb interactions @xcite . in this case , the most probable explanation for the metamagnetic behavior is the evolution of fm interaction due to the double exchange mechanism . here , the fm interaction evolves via the hund coupling between local moments and conduction electrons , while afm interaction can also arise via the exchange interaction between local moments @xcite . this would also result in the situation where both the fm and the afm interactions are important . although our results do not clarify whether the @xmath6-electron state in liv@xmath0o@xmath1 is itinerant or localized , the absence of metamagnetic behavior in liv@xmath0o@xmath2 seems to be of particular importance . the reduced @xmath21 in this system implies that the fm interaction is rapidly reduced by a slight disorder or defects . this may be a hint for the origin of the fm interaction in liv@xmath0o@xmath1 . since the @xmath6-electrons in liv@xmath0o@xmath1 involve degrees of freedom of orbital and charge quanta , orbital and/or charge order would also be suppressed by the geometrical frustration . this multiply frustrated state can be easily lifted by a small defect . the fm correlation in liv@xmath0o@xmath1 may be a result of the competition of the multiply frustrated state . to elucidate the electronic state and the mechanism for the evolution of the fm correlation in liv@xmath0o@xmath1 , more systematic experiments as well as theoretical studies should be performed . we have reported a metamagnetic behavior in liv@xmath0o@xmath1 around @xmath28 = 38 t and absence of metamagnetism in liv@xmath0o@xmath2 up to 40 t. arrot plot of the former shows the characteristic curve for a metamagnetic behavior , while that for the latter has a positive curve . this implies that metamagnetic behavior would not appear for liv@xmath0o@xmath2 even at much higher fields . moreover , high - field magnetic moment of this system is significantly smaller from that of liv@xmath0o@xmath1 . this qualitative difference in the magnetization of the two samples is of particular importance to consider the mechanism of the metamagnetic behavior . metamagnetic behavior out of a paramagnetic state has been observed in two cases : kondo - lattice systems and itinerant - electron systems with strong ferromagnetic spin - fluctuations . for the case of kondo - lattice systems , the high - field state is described by localized electrons with single - site kondo fluctuations . consequently , the magnetic moments at high fields are insensitive to a slight disorder or defects for kondo - lattice systems . for the present case , the significant difference in the magnetization of liv@xmath0o@xmath1 and liv@xmath0o@xmath2 does not agree with the local moment state . the kondo - lattice model is therefore unlikely as far as the mechanism for the metamagnetic behavior is concerned . instead , these results favor the evolution of intersite ferromagnetic spin - correlation at high fields for the origin of metamagnetic behavior in liv@xmath0o@xmath1 . this may lead to the interpretation that liv@xmath0o@xmath1 is a nearly - ferromagnetic metals , as is suggested by fujiwara _ et al . _ from their @xmath11li - nmr experiments in fact , many itinerant - electron systems with strong ferromagnetic spin - fluctuations exhibit metamagnetic behavior by fields @xcite . for liv@xmath0o@xmath1 , however , neutron scattering experiments have revealed the presence of antiferromagnetic fluctuations at low temperature @xcite . from these facts , we conclude that both the ferromagnetic and the antiferromagnetic interactions are important in liv@xmath0o@xmath1 , and the former is strongly enhanced by external fields . microscopic origin of the ferromagnetic correlation is unclear . the qualitative difference in the high - field magnetizations of liv@xmath0o@xmath1 and liv@xmath0o@xmath2 is likely to suggest that the effect of geometrical frustration for spin , orbital , and charge order is relevant to the ferromagnetic interaction . the other possibility for the ferromagnetic correlation would be the double exchange mechanism . 99 kondo s , johnston d c , swenson c a , borsa f , mahajan a v , miller l l , gu t , goldman a i , maple m b , gajewski d a , freeman e j , dilley n r , dickey r p , merrin j , kojima k , luke g m , uemura y j , chmaissem o and jorgensen j d 1997 _ phys . lett . _ * 78 * 3729 onoda m , imai h , amako y and nagasawa h 1997 _ phys . rev . _ b * 56 * 3760 ueda y , fujiwara n and yasuoka h 1997 _ j. phys . jpn . _ * 66 * 778 urano c , nohara m , kondo s , sakai f , takagi h , shiraki t and okubo t 2000 _ phys . * 85 * 1052 fujiwara n , ueda y and yasuoka h 1997 _ physica _ b * 237 - 238 * 59 + fujiwara n , yasuoka h and ueda y 1998 _ phys . rev . _ b * 57 * 3539 + fujiwara n , yasuoka h and ueda y 1999 _ phys . _ b * 59 * 6294 mahajan a v , sala r , lee e , borsa f , kondo s and johnston d c 1998 _ phys . _ b * 57 * 8890 kawakami k , sakai y and tsuda n 1986 _ j. phys . jpn . _ * 55 * 3174 muhtar , takagi f , kawakami k and tsuda n 1988 _ j. phys . jpn . _ * 57 * 3119 krimmel a , loidl a , klemm m , horn s and schober h 1999 _ phys . lett . _ * 82 * 2919 + murani a p 2000 _ phys . lett . _ * 85 * 3981 + krimmel _ et al . _ 2000 _ phys lett . _ * 85 * 3982 lacroix c 2001 _ can . j. phys . _ * 79 * 1469 nakajima y , amamiya y , ohnishi k , terasaki i , maeda a and uchinokura k 1991 _ physica _ c * 185 - 189 * 719 johnston d c 2000 _ physica _ b * 281 - 282 * 21 johnston d c , swenson c a and kondo s 1999 _ phys . _ b * 59 * 2627 kaps h , brando m , trinkl w , bttgen n , loidl a , scheidt e - w , klemm m and horn s 2001 _ j. phys . : condens . matter _ * 13 * 8497 merrin j , fudamoto y , kojima k m , larkin m , luke g m , nachumi b , uemura y j , kondo s and johnston d c 1998 _ j. mag . mag . mater . _ * 177 - 181 * 799 chmaissem o , jorgensen j d , kondo s and johnston d c 1997 _ phys . lett . _ * 79 * 4866 flouquet j , kambe s , regnault l p , haen p , brison j p , lapierre f and lejay p 1995 _ physica _ b * 215 * 77 schrder a , schlager h g and lhneysen h v 1992 _ j. mag . mag . mater . _ * 108 * 47 lhneysen h v , schlager h g and schrder a 1993 _ physica _ b * 186 * 590 hewson a c and rasul j w 1983 _ j. phys . _ c * 16 * 6799 for a review , goto t , fukamichi k and yamada h 2001 _ physica _ b * 300 * 167 sakakibara t , goto t , yoshimura k and fukamichi k 1990 _ j. phys . : condens . _ * 2 * 3381 monod p , felner i , chouteau g and shaltiel d 1980 _ j. physique _ * 41 * l511 rossat - mignod j , regnault l p , jacoud j l , vettier c , lejay p , flouquet j , walker e , jaccard d and amato a 1988 _ j. mag . mag . mater . _ * 76 - 77 * 376 takashita m , aoki h , terashimra t , uji s , maezawa k , settai r and nuki y 1996 _ j. phys . jpn . _ * 65 * 515 matsuhira k , sakakibara t , amitsuka h , tenya k , kamishima k , goto t and kido g 1997 _ j. phys . _ * 66 * 2851 sugawara h , namiki t , yuasa s , matsuda t d , aoki y , sato h , mushnikov n , hane s and goto t 2000 _ physica _ b * 281 - 282 * 69 mattens w c m , de chtel p f , moleman a c and de boer f r 1979 _ physica _ b * 96 * 138 moriya t 1985 _ spin fluctuations in itinerant electron magnetism _ , springer - verlag goto t , shindo y , takahashi h and ogawa s 1997 _ phys . rev . 14019 perry r s , galvin l m , grigera s a , capogna l , schofield a j , mackenzie a p , chiao m , julian s r , ikeda s i , nakatsuji s , maeno y and pfleiderer c 2001 _ phys . rev _ * 86 * 2661 ikeda s i , maeno y , nakatsuji s , kosaka m and uwatoko y 2000 _ phys . _ b * 62 * r6089 lee s - h , qiu y , broholm c , ueda y and rush j j 2001 _ phys . lett . _ * 86 * 5554 shiga m , fujisawa k and wada h 1993 _ j. phys . * 62 * 1329 nakamura h , yoshimoto k , shiga m , nishi m and kakurai k 1997 _ j. phys . : condens . matter _ * 9 * 4701 ballou r , lelivre - berna e and fk b 1996 _ phys . rev _ * 76 * 2125 kitaoka y and yasuoka h 1980 _ j. phys . jpn . _ * 48 * 1460 eyert v , hck k - h , horn s , loidl a and riseborough p s 1999 _ europhys . _ * 46 * 762 polatsek g and zevin v 1988 _ j. mag . mag . mater . _ * 73 * 205 anisimov v i , korotin m a , zlfl m , pruschke t , le hur k and rice t m 1999 _ phys . rev . lett . _ * 83 * 364

metamagnetic behavior has been observed in liv@xmath0o@xmath1 powder sample around 38 t at 4.2 k. on the other hand , magnetization for oxygen deficient liv@xmath0o@xmath2 shows no indication of metamagnetism up to 40 t , and shows substantially reduced magnetic moment compared to that of liv@xmath0o@xmath1 . these results suggest that ferromagnetic interaction is strongly enhanced by magnetic fields in liv@xmath0o@xmath1 , whereas antiferromagnetic interaction is dominant in liv@xmath0o@xmath2 .

the physical properties of galaxy nuclei are of larger interest in light of growing indications that evolution in the central few parsecs is linked to the global properties of the surrounding galaxy . this connection is evident particularly in correlations between the mass of a central black hole , and bulge mass ( e.g. , kormendy & richstone 1995 ; hring & rix 2004 ) or velocity dispersion ( gebhardt et al . 2000 ; ferrarese & merritt 2000 ) . star formation in the centers of galaxies is also known to take on unusual forms that may be in symbiosis with a central mass concentration , or the larger structure of the galaxy ( i.e. , bulge or bar ) . a rich phenomenology is seen in the central few parsecs of galaxy nuclei within the local group , but our knowledge of nuclei at greater distances is limited by a corresponding reduction in spatial resolution . this problem has particularly afflicted ground - based optical spectroscopic studies , that typically sample nuclei with apertures encompassing hundreds of parsecs or more . to investigate the structure and constituents of galaxy nuclei on small scales , we therefore undertook a survey of nearby nuclei with stis ( sunns ) , a program of _ hubble space telescope _ ( _ hst _ ) , narrow - slit spectroscopy for a sample of nearby galaxies outside the local group . the target objects were selected on the basis of nebular emission in their centers , and are thus `` active '' in some sense , although at low levels that are very common in bright galaxies . the sunns data are appropriate for addressing several major questions , as described in part in our previous papers in this series . the long - slit measurements of gas kinematics can be used to measure or place limits on the mass of a central black hole ( sarzi et al . 2001 , 2002 ) . the spectra coupled with _ hst _ imaging can be used further to probe the structure of nuclear gas disks , which we find to be influenced in many cases by nongravitational effects ( ho et al . the use of small apertures is very helpful in isolating nuclear emission components that are lost in the contamination by circumnuclear sources , as demonstrated by our discovery of unusual broad - line emission in two galaxies ( shields et al . 2000 ; ho et al . 2000 ) . small apertures are also important for obtaining a description of the stellar population and star formation history within the central tens of parsecs ( sarzi et al . 2005 ) . in the current paper we present a summary description of the sunns experimental design , and focus on new results for the nuclear emission - line characteristics of the sample . the spatial distribution and excitation properties of nebular emission can be used to draw conclusions about the dominant energy sources in galaxy centers , and as part of our investigation we explore the significance of aperture effects in determining the spectroscopic appearance of galaxy nuclei . understanding the energetics of emission - line nuclei is of fundamental importance for tracing rates of star formation and black hole growth in these low - power but nearly ubiquitous objects . the data reported here also provide an improved basis for comparing energetic phenomena and structure in the milky way s nucleus with the centers of other similar galaxies . _ hst _ observations ( go-7361 ) were obtained for 24 galaxies within 17 mpc of the milky way , drawn from the palomar spectroscopic survey of nearby galaxies ( filippenko & sargent 1985 ; ho , filippenko & sargent 1997a ) . the sample was chosen to be distance - limited to maximize spatial resolution , and was selected to have nebular emission potentially suitable for kinematic studies and energetic probes . we focused our study on disk galaxies , since these objects most commonly show nebular emission in their centers . we restricted the sample to hubble types s0 sbc , since the nuclei of later types are more likely to be weakly defined , affected by dust , or otherwise subject to confusion . the sample was limited to galaxies with inclinations @xmath2 to minimize extinction by dust residing in the disk . we selected objects with h@xmath1 or [ ] @xmath36583 line emission @xmath4 ergs s@xmath5 @xmath6 within a @xmath7 aperture ( from ho et al . 1997a ) , with spectroscopic classifications fairly evenly distributed among seyfert , liner 1 , liner 2 , liner/ `` transition , '' and categories . the result is listed in table 1 . data were acquired in 1998 and 1999 with the space telescope imaging spectrograph ( stis ) on board _ hst _ , using the 02 @xmath8 50 aperture . for each galaxy , acquisition images were obtained using the f28x50lp long - pass filter with 30 s integration , and peak - up exposures were taken in order to center the spectrograph slit on the target nucleus . inspection of these and other archival _ hst _ images indicates that in all but one case , the aperture was well - centered on an obvious brightness peak at the morphological center , with no indication of confusion by dust obscuration or other structure ; errors in aperture centering due to dust effects are @xmath9 pixels @xmath10 . the exception occurred in ngc 4138 , where the peak - up procedure positioned the slit on a bright foreground star offset @xmath11 from the nucleus ; we consequently dropped this object from our analysis . the slit position angle ( pa ) was left unconstrained and thus was set by the spacecraft roll angle at the time of observation . for a given galaxy , all exposures were obtained at a single pa . two - dimensional ( 2-d ) spectra were obtained with the g430l and g750 m gratings , yielding spectra spanning 33005700 and 63006850 with fwhm spectral resolution for extended sources of 10.9 and 2.2 , respectively . these intervals were chosen in order to include prominent emission lines , and to measure continuum intervals diagnostic of stellar populations and possible active galactic nucleus ( agn ) contributions . the medium - resolution grating was selected for the red setting so as to enable high precision velocity measurements of the h@xmath1 @xmath36563 and [ ] @xmath126548 , 6583 lines , which tend to be the strongest optical emission features , and hence optimal for kinematic studies . two exposures totaling @xmath13 s and three exposures totaling @xmath14 s were obtained with the blue and red settings , respectively . for most of the target galaxies , the telescope pointing was adjusted between duplicate exposures so as to shift the location of the nucleus along the slit by increments of several pixels , to facilitate removal of hot pixels . the resulting data were calibrated using standard methods . the 2-d spectra were bias- and dark - subtracted , flat - fielded , aligned , and combined into single frames . cosmic rays and hot pixels remaining in the combined 2-d spectra were cleaned according to the prescription of rudnick , rix , & kennicutt ( 2000 ) , and the spectra were then corrected for geometrical distortion and wavelength and flux calibrated with standard stsdas procedures within iraf . during each reduction step , care was taken to correctly propagate the information contained in the error array corresponding to each galaxy 2-d spectrum and initially provided by the stis pipeline . to represent the nuclear spectrum of each galaxy , we extracted aperture spectra five pixels wide ( @xmath15 ) , centered on the brightest part of the continuum , using the iraf task apall . the extracted spectra thus consist of the central emission convolved with the stis spatial point - spread function and sampled over an aperture of @xmath16 , roughly corresponding to a circular aperture with radius @xmath17 , or 8.2 pc for the mean sample distance of @xmath1813 mpc ( or @xmath19 pc for the full range of galaxy distances ) . in addition , from the 2-d error array corresponding to each 2-d spectrum , we extracted the central five , one - pixel - wide aperture spectra , and added them quadratically to obtain the error array corresponding to each nuclear extraction . the resulting 1-d spectra for the blue ( g430l ) setting appear in sarzi et al . ( 2005 ) , while those for the red ( g750 m ) setting are shown here in figure 1 . the emission - line equivalent widths in the nuclear spectra are often small and consequently require subtraction of the underlying starlight in order to measure accurate emission - line fluxes . we modeled the continuum with linear combinations of stellar spectra from the miles library of sanchez - blazquez et al . ( 2006 ) , with weights and velocity broadening optimized with the direct - fitting method described in sarzi et al . this is similar to the method of rix & white ( 1992 ) , except that the spectral regions affected by nebular emission are no longer excluded from the fitting process , since the emission - lines are treated as gaussian templates and fitted simultaneously with the stellar templates to the observed spectra . this has the advantage of maximising the spectral information available to the fitting algorithm . as the blue and red spectra have different spectral resolutions they can not be fitted simultaneously . on the other hand , the presence of stronger emission lines and the limited number of absorption features makes it difficult to independently match the stellar continuum of the red bandpass . we therefore fit first the blue spectra , which are important for constraining the mix of template stellar types , and use the resulting optimal combination of templates to match the red spectra , allowing only for a different velocity broadening . the fitting of the blue spectra include reddening due to interstellar dust , both in the milky way and in the sample galaxies , and due to dust in the emission - line regions . the latter affects only the fluxes of the emission - line templates , and is constrained by the expected decrement of the balmer lines . in matching the blue spectra we imposed the same kinematics for both forbidden and recombination lines . we also restricted the stellar template library to spectra with values for the h@xmath20 absorption line - strength index exceeding 0.1 . in the present study we are interested in obtaining the best empirical fit to the starlight , and no a priori assumptions are made concerning the relative weighting of different stellar types ; our method here consequently differs somewhat from that employed in our population analysis ( sarzi et al . 2005 ) which is based on templates corresponding to parametrized star formation histories . the fits to the blue spectra are very similar to the parametrized fits plotted in sarzi et al ( 2005 ) , although the miles spectra can match several spectral regions where models such as those of bruzual & charlot ( 2003 ) would need to account for non - solar abundance of @xmath1-elements before they can match the data . the best description for the stellar continuum in the red bandpass is overplotted on the spectra shown in figure 1 . the fits for the red spectra differ from those for the blue bandpass not only in that they use a different velocity broadening and the optimal combination of templates from the fit on the blue spectra , but also in that they employ a generalized multiplicative second - order polynomial adjustment instead of a specific dust reddening law . emission - line templates were used also while fitting the red spectra , although for galaxies with complex line profiles or broad components , the regions affected by emission were simply excluded from the fit . for many of our target objects , the emission lines within the nuclear aperture are weak . we attempted to measure lines that were visible in much of the sample on inspection , or likely to be present based on the behavior of line - intensity ratios in other nebular sources . specifically , we measured or placed limits on the flux in [ ] @xmath33727 , h@xmath21 @xmath34340 , h@xmath20 @xmath34861 , [ ] @xmath124959 , 5007 , [ ] @xmath36300 , [ ] @xmath126548 , 6583 , h@xmath1 @xmath36563 , and [ ] @xmath126716 , 6731 . the lines were measured from the starlight - subtracted spectra using specfit ( kriss 1994 ) as implemented in iraf . this routine is suitable for fitting complicated line profiles with additional contraints between lines , while also generating error estimates for measured quantities . gaussian profiles were assumed for the lines ; a few sources required double gaussians to obtain a good match to the observed lines , and additional components were included in cases where there was evidence of broad balmer emission ( 3.3 ) . to improve the fitting accuracy , we introduced constraints to reduce the number of fit parameters when justified by physical arguments , or by our previous experience with ground - based spectra ( e.g. , ho et al . the h@xmath1 , [ ] , and [ ] lines were assumed to share a common centroid velocity , and the [ ] and [ ] lines were assumed to share a common velocity width . a flux ratio of 1:3 was assumed for the [ ] @xmath126548 , 6583 and [ ] @xmath124959 , 5007 doublets , as dictated by atomic physics ( e.g. , osterbrock 1989 ) . the h@xmath20 and h@xmath21 lines were also assumed to share a common velocity and profile . the g430l and g750 m spectra were fit separately because of their differing spectral resolution and possible zero - point wavelength offsets . in determining relative wavelengths for lines with shared velocities , rest wavelengths to 6 significant figures were taken from kaufman & sugar ( 1986 ) for the forbidden lines and from wiese , smith , & glennon ( 1966 ) for the balmer lines . the resulting line fluxes are listed in table 2 . the line fluxes are uncertain due to both statistical fluctuations and systematic errors . specfit uses a @xmath22 minimization algorithm , which provides one estimate of the line flux uncertainties . as a measure of the possible systematic error , we also calculated an equivalent 1@xmath23 gaussian uncertainty based on the standard deviation @xmath24 per pixel in the residual continuum after starlight subtraction . the standard deviation was computed in regions devoid of emission lines and traces structure due to imperfect continuum subtraction , in addition to statistical fluctuations . the corresponding error is then given by a gaussian profile with amplitude of 1@xmath24 , and a width typical of the measured narrow lines ( fwhm of 300 km s@xmath5 for g750 m and 550 km s@xmath5 for g430l , where the width includes instrumental broadening ) . in determining a final value for the flux uncertainty , we adopted the largest of 1 ) the specfit error , 2 ) the gaussian 1@xmath23 uncertainty , and 3 ) a conservative minimum error of 10% of the line flux . as a consistency test , we repeated the line measurements after subtracting the best - fit multi - age stellar population models for these spectra that are described by sarzi et al . line fluxes were measured from the continuum - subtracted spectra by direct integration over the line profile , using splot in iraf . the results showed good consistency within the measurement uncertainties with the values listed in table 2 . the _ hst _ aperture employed for this study is much smaller than typical ground - based apertures , and a comparison of nebular properties as seen on these different scales provides useful information on the structure of emission nuclei . the palomar spectra used to define the current sample employed a @xmath25 aperture , which is @xmath26 larger in linear dimension and @xmath27 larger in area than the sunns nuclear aperture . the linear scales probed by these data are @xmath28 vs @xmath29 pc . the degree to which the nebular emission in our sample galaxies is centrally concentrated can be gauged from a comparison of fluxes measured through the sunns and palomar apertures . the result is shown in figure 2 , which displays the ratio of narrow h@xmath1 fluxes as a function of palomar spectroscopic class . the sequence of categories on the abscissa is intended to reflect the degree to which the nuclei show evidence of being powered by accretion . in this and subsequent figures displaying ratios , the plotted error bars are propagated from the uncertainties listed in table 2 for the sunns measurements and in ho et al . ( 1997a ) for the palomar data . for a uniform nebular surface brightness , the flux ratio should be @xmath30 dex . in the majority of cases the ratio is measured to be higher , indicating some degree of central concentration . on the other hand , for all of the galaxies , the small - aperture measurement shows a significant reduction in flux relative to the palomar value , indicating that the central nebulae in these objects are resolved at _ hst _ resolution . figure 2 shows evidence of a trend in that sources showing the strongest indications of accretion power ( liner 1s and seyferts ) display the largest ratios , implying greater concentration . in contrast , the nuclei show the smallest ratios , and in fact four of the five nuclei are undetected in h@xmath1 in the sunns aperture . the star formation powering `` nuclei '' apparently resides on scales of tens of parsecs or more ( see , e.g. , hughes et al . 2003 , cid fernandes et al . 2004 , and gonzlez delgado et al . 2004 for related results ) . the narrow emission - line equivalent widths ( ews ; see table 2 and figure 3 ) show a similar trend , with larger ew(h@xmath1 ) found in the sunns aperture for the liner 1 and seyfert nuclei , but smaller ew(h@xmath1 ) in the small aperture for the majority of , transition , and liner 2 nuclei . while a direct measurement of line ratio gradients is possible in principle from the _ hst _ long - slit data , in practice the signal - to - noise ratio severely hampers detailed study of spatial variations . comparison of line ratios measured in the sunns and palomar apertures provides an alternative means of quantifying excitation gradients . in addition to yielding a factor of @xmath31 more signal , the palomar aperture avoids possible problems with azimuthal variations in extended emission that are not adequately sampled by the narrow sunns slit . line ratios useful as diagnostics that are also measured for a significant number of sample members are limited , but four particularly relevant ratios are [ ] @xmath35007/h@xmath20 , [ ] @xmath36300/h@xmath1 , [ ] @xmath36583/h@xmath1 , and [ ] @xmath126716 , 6731/h@xmath1 . these diagnostics are commonly employed in classification of nebular sources ( e.g. , veilleux & osterbrock 1987 ) , and standard diagrams used for this purpose are shown in figure 4 , with measurements from both palomar and sunns plotted . there is considerable variation in the behavior of different objects in figure 4 , but several conclusions can nonetheless be drawn from these plots . first , the line ratios are often rather insensitive to aperture size , and in the present comparison show changes that are usually less than a factor of two . related to this , the sources generally do not shift their position by large amounts in the line ratio diagrams , which means that their classification is insensitive to aperture size . this finding has interesting and important implications that we examine in detail below . one limitation of the two - dimensional line ratio diagrams is that the sunns observations provide only upper limits for both [ ] and h@xmath20 for nearly half the sample , thus preventing their inclusion in figure 4 . an alternative is to compare line intensity ratios individually , and for this purpose we plot in figure 5 the `` ratio of ratios '' for the measurements obtained in the two apertures . several patterns are discernable in figure 5 . [ ] /h@xmath1 is arguably the best indicator of excitation beyond the contributions of normal regions , since it requires a significant partially ionized zone for its production . the [ ] line is in fact most significant in the liner 1 and seyfert nuclei , and figure 5 shows a clear tendency for the ratio to be larger in the small aperture for these sources . a weaker trend in the same direction exists for [ ] /h@xmath1 , but no trend is evident for [ ] /h@xmath1 . the more limited measurements of [ ] /h@xmath20 show little dependence on aperture size , except for some tendency for the ratio to decrease in the small aperture in the seyfert nuclei . density gradients may be a factor in the nebular emission structure reflected in figure 5 . previous studies have shown indirect evidence from line width - critical density correlations ( e.g. , pelat , alloin , & fosbury 1981 ; filippenko & halpern 1984 ) as well as direct evidence from the [ ] density diagnostic ( barth et al . 2001 ) that densities increase toward small radii in the narrow - line regions of agns . figure 6 shows the density - sensitive [ ] @xmath36716/@xmath36731 line ratio for the sample sources as measured in both the palomar and sunns spectra . the ratio is lower in all but one case for the small aperture measurement , implicating a characteristically higher density on smaller radial scales ; median values are 1.2 from the palomar measurements and 0.95 from sunns , corresponding to densities of @xmath32 @xmath33 and @xmath34 @xmath33 , respectively . these values are presumably emissivity - weighted averages of a potentially much broader span of densities . the different spectral classes show substantial overlap in their distributions in figure 6 . the different behavior observed in the [ ] , [ ] , and [ ] lines in figure 5 can potentially be understood as a consequence of the varied critical densities for the forbidden lines . in the sunns aperture a significant fraction of the nebular gas for the liners and seyferts evidently has a density similar to or greater than the [ ] @xmath126716 , 6731 critical densities @xmath35 of @xmath36 @xmath33 and @xmath37 @xmath33 ( e.g. , peterson 1997 ) , which could act to depress the [ ] /h@xmath1 ratio in the small aperture . the [ ] @xmath35 values are the lowest of those for the lines shown in figure 5 . in contrast , [ ] @xmath36300 , with @xmath38 @xmath33 , remains strong in the small aperture ; the larger [ ] /h@xmath1 ratios found with sunns may reflect an increased importance of partially ionized gas on small scales , but would also be consistent with a higher nebular temperature in dense regions where cooling by low-@xmath35 emission ( e.g. , from infrared fine - structure lines ) is inefficient . [ ] @xmath36583 , with an intermediate @xmath35 value of @xmath39 @xmath33 , shows a weaker trend in the same direction . the overall results suggest that a significant fraction of the nebular gas in galaxies on scales of @xmath29 pc has densities of @xmath40 @xmath33 . the [ ] @xmath35007 line , with @xmath41 @xmath33 , does not readily fit into this scenario . for the seyfert nuclei in particular , the [ ] /h@xmath20 ratio tends to decrease in the small aperture , and there is some degree of negative correlation between the [ ] /h@xmath20 and [ ] /h@xmath1 ratios , suggesting that ionization structure which will be sensitive to ionization parameter and hence the detailed radial variation in density also plays a role in the aperture dependence . the existing palomar observations provided evidence that six of our 23 target galaxies exhibit weak quasar - like broad h@xmath1 emission , resulting in classification as liner 1.9 ( ngc 2787 , ngc 4143 , ngc 4203 , ngc 4450 ) or seyfert 1.9 ( ngc 3982 , ngc 4501 ) nuclei ( ho et al . we investigated whether a broad component was also present in the _ hst _ spectra for these sources and the other objects in our sample . the sunns data for two of the liner 1.9 galaxies , ngc 4203 and ngc 4450 , display evidence of not only a gaussian broad h@xmath1 line , as seen from the ground , but also a very broad , double - shouldered component ( see fig . 1 ) ; these findings are discussed in detail elsewhere ( shields et al . 2000 ; ho et al . 2000 ) . we attribute the discovery of the additional broad features to the improved contrast in the small _ hst _ aperture between the broad emission , which presumably arises within @xmath42 pc of the center , and the stellar continuum plus narrow - line emission , which is more widely dispersed in origin . the implication is that such broad emission may be common in low - luminosity agns , and a signature of the accretion structure in low accretion - rate nuclei . other examples of double - shouldered emission profiles found in _ hst _ spectra of liner 1 nuclei are described by bower et al . ( 1996 ) and barth et al . ( 2001 ) . figure 7 shows the results of a profile decomposition using the sunns data for the remaining four galaxies where broad emission was previously suspected . a common profile shape was employed to represent the [ ] , narrow h@xmath1 , and [ ] lines ; double gaussians were adopted to better match the profiles for ngc 3982 and ngc 4143 . an additional broad h@xmath1 component with unconstrained width and central wavelength was included in the fit . strong evidence for a broad emission feature is found in ngc 2787 and ngc 4143 , while marginal evidence for such emission is present in ngc 4501 , and a reasonable fit for ngc 3982 can be obtained with no broad component . the narrow - line fluxes resulting from this decomposition are listed in table 2 ; broad - line fluxes and the fraction @xmath43 of the total h@xmath1 + [ ] complex represented by the broad h@xmath1 feature are listed in table 3 , along with the corresponding numbers measured in the palomar spectra . for the four galaxies where broad emission appears to be confirmed ( ngc 2787 , ngc 4143 , ngc 4203 , and ngc 4450 ) , @xmath43 increases if these values are taken at face value , as expected if the narrow emission is resolved and diminished in the small aperture while the broad emission remains unaffected by slit losses . closer scrutiny of the deblending results suggests a problem with the fitting procedure , however . the values listed in table 2 imply h@xmath1/h@xmath20 ratios for the narrow component of ( @xmath44 , @xmath45 , @xmath46 , @xmath47 ) for ngc 2787 , ngc 4143 , ngc 4203 , and ngc 4450 , respectively . these ratios are considerably smaller than the theoretical ratio of 3.1 predicted for agn narrow - line regions ( osterbrock 1989 ) , and are impossible to understand on physical grounds . the low values are unlikely to be caused by broad - line contamination of h@xmath20 ; ngc 4203 and ngc 4450 both show evidence of broad emission in h@xmath20 and h@xmath21 , but this component is included as a separate component when determining the narrow balmer fluxes . no evidence of broad emission is seen in these lines for ngc 2787 and ngc 4143 . the narrow h@xmath21/h@xmath20 ratios are , moreover , normal : ngc 2787 , ngc 4143 , ngc 4203 , and ngc 4450 display h@xmath21/h@xmath20 = @xmath48 , @xmath49 , @xmath50 , and @xmath51 , respectively , in reasonable agreement with the predicted recombination value of @xmath52 . there is no evidence of significant discontinuities between the g430l and g750 m spectra indicative of flux calibration errors ; almost all of the other objects in our sample show narrow h@xmath1/h@xmath20 ratios consistent with the theoretical prediction ( ngc 4698 has a ratio of @xmath53 ) , or a larger ratio attributable to reddening . it is noteworthy that for the four galaxies in question , the forbidden line / h@xmath1 flux ratios are unusually large for liners or seyfert nuclei ( although it must be kept in mind that the sunns aperture does not sample the full narrow - line region ) . given these findings , we strongly suspect that the deblending procedure outlined above and illustrated in figure 7 underestimates the true narrow h@xmath1 flux which in turn would produce an overestimate of the broad component flux . we experimented with alternative parameterizations of the fits for ngc 2787 and ngc 4143 , including decoupling of the narrow h@xmath1 line widths from the other lines , forcing the narrow h@xmath1 flux to match the value predicted from h@xmath20 , and forcing the narrow h@xmath1 width to match that of h@xmath20 . in general , these approaches were not particularly satisfactory at matching the detailed profile structure , resulting in residuals larger than seen in figure 7 . this outcome suggests that the narrow and/or broad profiles are rather more complex than simple gaussians and that significant differences may exist between the different narrow - line profiles . a possible broader component of [ ] than is currently represented in our deblending model may also be suggested by the redward displacement of the broad h@xmath1 feature relative to the narrow emission found in the four galaxies at issue here , since the stronger line in the doublet , @xmath36583 , would thereby add flux that our default model would interpret as redshifted h@xmath1 emission . we therefore caution that the fluxes of narrow h@xmath1 and possibly [ ] in table 2 for these objects may suffer from systematic errors in excess of the listed uncertainties . for purposes of comparison with the palomar results , however , the fact that we employed a very similar fitting algorithm may mitigate the influence of these errors . it is noteworthy that the palomar measurements from ho et al . ( 1997a ) yield h@xmath1/h@xmath20 ratios of 1.9 , 1.6 , 1.0 , and 2.6 for ngc 2787 , ngc 4143 , ngc 4203 , and ngc 4450 , respectively , which likewise fall short of the recombination prediction and suggest similar systematic errors to what we have encountered with the sunns data . in ratioing the sunns measurements to the palomar measurements ( e.g. , in figures 2 , 3 , and 5 ) , we can thus expect these errors to cancel out at least partially . a best estimate of the true narrow h@xmath1 flux is probably that derived from the narrow h@xmath20 flux , scaled by the recombination prediction , which yields ( @xmath54 , @xmath55 , @xmath56 , @xmath55 ) @xmath57 ergs s@xmath5 @xmath6 for ngc 2787 , ngc 4143 , ngc 4203 , and ngc 4450 , respectively . comparison of the sunns broad h@xmath1 fluxes with those from palomar reveals a surprising trend , in that the sunns flux is in all cases smaller . taking the profile decomposition results at face value , the median reduction is a factor of @xmath58 . it is not clear how this outcome could result from systematic errors in the deblending process . we also do not expect the broad feature to suffer from aperture losses in the sunns data , since the emission is expected to arise from sub - parsec scales . for spatially unresolved sources the stis flux calibration corrects for aperture transmission losses to within an accuracy of @xmath59% at h@xmath1 ( bohlin & hartig 1998 ) . the present result could be understood physically if the broad emission is variable ; the sources in the palomar survey identified as l1.9 or s1.9 nuclei would then have preferentially been in a high flux state , since sensitivity limits dictate that objects in low flux states would not be detected as broad - line objects . if the sources vary on timescales shorter than the decade interval between the palomar and sunns observations , we would then expect that the broad h@xmath1 flux would be lower in a majority of cases in the second - epoch observation . we would also predict that some number of sources seen as type 2 nuclei in the palomar data would display new detectable broad components at the second epoch . variability in broad line emission would not be surprising and has been reported in similar low - luminosity agns previously ( e.g. , eracleous & halpern 2001 and references therein ) . for the sources in our sample that appear in the palomar spectra as narrow - line objects , it is therefore a surprise that the _ hst _ spectra provide no new indications of broad h@xmath1 . this conclusion was reached after examining the residuals from fitting with purely narrow - line profiles as described above . the lack of newly detected broad - line sources is curious with respect to the variability prediction noted above , but also given the expected increase in sensitivity in the small _ hst _ aperture due to the reduction in circumnuclear contamination . investigation of a larger sample would be desirable in order to clarify the generality of this result and its implications . an interesting question that can be addressed with the results of this study is the extent to which the spectroscopic classification of a galaxy nucleus depends on the measurement aperture employed . the results from 3.1 indicate that the sunns aperture typically excludes most of the emission that is seen in the palomar measurement , and that nebulosity that is clearly dominated by an accretion source tends to be the most centrally concentrated . we might therefore expect the _ aperture to isolate gas that is higher in excitation / ionization and more dominated by the central engine , to the extent that one is present . likewise , the small - aperture measurement should increase the probability of detection for broad h@xmath1 components . in summary , the simple prediction from these considerations is that the sunns nuclear spectra should appear more agn - like than the corresponding palomar spectra , for nuclei where accretion is important as a power source . objects of particular interest in this regard are the /liner transition nuclei . these sources are intermediate in their emission properties between nuclei and liners or seyfert nuclei ( e.g. , ho et al . 1997a ; hao et al . 2005 ) ; other labels applied to such objects are `` composite nuclei '' ( e.g. , kennicutt , keel , & blaha 1989 ; vron , gonalves , & vron - cetty 1997 ) and `` weak [ ] liners '' ( e.g. , filippenko & terlevich 1992 ) . the standard interpretation for these sources , which are very common ( @xmath60% of bright galaxies ; ho , filippenko , & sargent 1997c ) , is that their spectra trace a mix of agn emission and circumnuclear regions . this picture is supported by kinematically distinct emission for the two components in some galaxies ( vron et al . 1997 ) , and by aperture - dependent line ratios on kpc scales for agns that may reflect different admixtures of the two emission sources ( storchi - bergmann 1991 ) . under this scenario , we would expect the nuclei to more strongly resemble liners or seyferts in the _ hst _ spectra , since the small aperture should act to exclude a part of the circumnuclear nebulosity . from the results reported in 3 there is only limited support for these predictions . the liner 1s and seyfert nuclei show a modest aperture dependence in the line ratios that appears to reflect at least partially density effects ( 3.2 ) . to the extent that ionization gradients are present , it seems that the ionization often _ decreases _ on smaller scales . if we apply to the sunns measurements the same classification criteria as employed by ho et al . ( 1997a ) for the palomar observations , we find that only three objects formally change their classification . ngc 3992 and ngc 4548 , classified respectively as a transition object and a liner 2 in the palomar spectra , show higher excitation in the small aperture , with an /h@xmath20 ratio consistent with a borderline seyfert identification . conversely , ngc 4698 , with a palomar classification as a seyfert 2 , shows a decrease in excitation in the sunns data consistent with classification as a liner or transition nucleus . the composite model for transition objects is not strongly supported by these data . approximately half of the admittedly small original sample of transition sources shows little aperture dependence in the line ratios , and in particular these objects do not appear as pure liners or seyferts in the sunns measurement . none of the narrow - line objects reveal new evidence of broad h@xmath1 emission when viewed at _ hst _ resolution ( see also gonzlez delgado et al . 2004 ) . the behavior of the transition objects is a surprise . the lack of strong emission - line gradients in some sources may indicate that the length scale over which the photoionizing energy of an accretion source is deposited is comparable to that of a circumnuclear star - forming zone ( i.e. @xmath28 pc ) , so that the influence of both energy sources is spatially commingled . this would involve a different gas distribution or geometry than is typical of the liner 1s and seyferts , which tend to have highly concentrated nebulosity around an accretion source , and may have interesting implications for the available fuel supply . alternatively , a distributed ionization source , such as turbulent mixing layers ( e.g. , begelman & fabian 1990 ) or evolved stars ( e.g. binette et al . 1994 ) , rather than simply a central accretion source , may be important in these nuclei ; the magnetic field may also be energetically significant in these regions . a distributed source is potentially supported by our finding that transition and liner 2 nuclei in the sunns sample preferentially show evidence of a nuclear component of stars with ages @xmath42 gyr ( sarzi et al . 2005 ; see also gonzlez delgado et al . investigation of a larger sample of transition objects is important for defining the properties of these sources and determining their nature . the decrease in ionization with decreasing radius in the seyfert 2 nuclei is likewise unexpected , and scrutiny of larger samples is needed to understand this finding . integral - field spectroscopic observations could prove useful to understanding the role of different ionizing sources at different radii , as recently shown by mazzuca et al . ( 2006 ) in their study of the galaxy ngc 7742 , which hosts a circumnuclear starburst and an active nucleus of composite nature . while the sample studied here consists of nearby emission nuclei , there is one such source notably missing from our study , namely the milky way nucleus . integrating the galactic center into a larger context of galaxy nuclei is complicated observationally by the large extinction to this region ( @xmath61 mag ; rieke , rieke , & paul 1989 ) , and by its close proximity , which results in an angular scale ( 1@xmath62 0.1 pc ) substantially different from extragalactic sources . spectroscopic measurements obtained for both external galaxies and the galactic center on the _ same _ metric scale in the _ same _ bandpass remain very limited . the sunns data , however , offer a new basis for comparison by at least matching the approximate metric scale of many measurements of the galactic center . the milky way , like the sunns galaxies , is host to a low - luminosity emission nucleus . the rate of photoionizations by lyman continuum photons for the galactic center is tabulated as a function of radius by mezger , duschl , & zylka ( 1996 , @xmath63 in their table 6 ) , based on radio measurements of free - free emission . after rescaling to a distance to the galactic center of @xmath64 kpc , the results indicate @xmath65 s@xmath5 within @xmath66 pc and @xmath67 s@xmath5 within @xmath68 pc . the h@xmath1 luminosity @xmath69 emerging within @xmath70 pc predicted from recombination theory with negligible extinction would be @xmath71 ergs s@xmath5 . objects with this @xmath69 are common in the full palomar spectroscopic survey , which has a slightly lower median @xmath72 ergs s@xmath5 for an equivalent radius of @xmath73 pc ( i.e. , for a circular aperture with area equal to the rectangular palomar aperture ; ho et al . allowing for modest extinction of the galactic center emission and the somewhat different metric apertures would bring the milky way value closer to the palomar median . for comparison purposes with sunns , an estimate of @xmath63 within @xmath74 pc for the galaxy would be desirable , but is not readily available ; with the existing data we can nonetheless estimate that the ratio of recombination luminosities within @xmath75 and @xmath28 pc is approximately one order of magnitude , which is broadly consistent with our measured ratio for other galaxies ( figure 2 ) . the galactic center is thus a very typical nucleus in terms of its nebular luminosity and concentration . the sunns data provide a further basis for comparing interstellar conditions in the galactic center with those in other galaxy nuclei . nebular densities within @xmath29 pc tend to be rather high . as discussed in 3.2 , the [ ] diagnostic indicates a typical density of @xmath76 @xmath33 , but other lines point to the presence of gas with densities of @xmath77 @xmath33 in the sunns nuclei . very similar results are found in the galactic center , using constraints from infrared fine - structure line ratios and emission measures ( e.g. , genzel , hollenbach , & townes 1994 ; beckert et al . 1996 , and references therein ) . the _ hst _ spectra do not provide quantitative constraints on the gas - phase abundances , although our analysis of the _ stellar _ abundances with these data indicates that typical metallicities @xmath78 z@xmath79 ( sarzi et al . 2005 ) . this range matches estimates of the nebular metallicity of the galactic center ( shields & ferland 1994 ) , and estimates of metallicity for stars in the central parsecs also suggest a typical value of @xmath80 z@xmath79 ( ramrez et al . 2000 ; blum et al . 2003 ) . although we can not see the optical spectrum of the galactic center , we can make some inferences about its optical spectral class from the results noted above and other measurements . the galactic center provides compelling evidence for the presence of a black hole with a mass @xmath81 m@xmath79 ( e.g. , schdel et al . 2003 ; ghez 2004 ) ; this mass and the stellar velocity dispersion @xmath82 for the milk way bulge are in reasonable accord with the @xmath83 correlation defined by other galaxies . the central black hole is apparently accreting , based on its appearance as the compact radio source sgr a@xmath84 ( balick & brown 1974 ) and an associated x - ray source ( baganoff et al . the accretion luminosity is , however , very feeble ( @xmath85 ergs s@xmath5 ; narayan et al . 1998 , and references therein ) , and thus insufficient to power a significant fraction of the nebular emission . the nebular emission on scales from @xmath86 pc ( e.g. , lutz et al . 1996 ) to @xmath87 pc ( e.g. , simpson et al . 1999 ) is characterized by low - ionization lines . the lack of both a strong compact x - ray source and high - ionization nebulosity implies that the milky way is not a seyfert nucleus . likewise , the central parsec of the galaxy does not exhibit strong , quasar - like permitted lines from high - velocity gas ; the galactic center is not a broad - line ( type 1 ) nucleus . would our counterparts in other galaxies regard the milky way as any sort of agn , of the types found in the palomar spectroscopic survey ? if so , our galaxy would have to be a liner or /liner transition object . a liner generated via photoionization is excluded , based on the absence of a significant compact hard x - ray source to power the nebulosity , as well as from detailed aspects of the infrared and radio emission lines ( shields & ferland 1994 ; lutz et al . 1996 ) . recent observations with _ reveal diffuse hard x - ray emission from the central @xmath88 pc , with a luminosity of @xmath89 ergs s@xmath5 ( muno et al . 2004 ) . while insufficient to power significant nebular emission directly , the hard x - rays are of uncertain origin , and would require a kinetic energy input of @xmath90 ergs s@xmath5 if generated in ( unbound ) thermal plasma ( see muno et al . 2004 for details ) . if a fraction of this mechanical energy were transferred to the nebular plasma , it could modify the excitation of the gas , resulting in enhancement of optical collisionally excited lines . however , the nebular gas is clearly not heavily shocked ; the plasma is not highly ionized , and at least within @xmath91 pc of the center , the electron temperature as revealed by radio line - to - continuum ratios is only @xmath92 k ( roberts & goss 1993 ) . while this value may be slightly higher than expected from simple stellar photoionization ( shields & ferland 1994 ) , it appears that the gas is not heated sufficiently to power optical forbidden lines at the level seen in liners . we conclude that the center of the milky way would appear spectroscopically as an nucleus or possibly a transition object for an extragalactic optical observer . as in most of the sunns galaxies ( sarzi et al . 2005 ) , the inner stellar bulge of our galaxy is predominantly old ( blum et al . 2003 ; van loon et al . 2003 ; although cf . figer et al . 2004 ) , but some star formation has occurred in recent episodes , and in the case of the milky way , the ultraviolet radiation from normal massive stars apparently dominates the nebular ionization . for galaxies like our own , with a hubble type of approximately sbc , the palomar spectroscopic survey found that the fractions hosting and transition nuclei are @xmath93% and @xmath94% , respectively ( ho et al . 1997c ) . in its nebular excitation , star formation history , and relation to the larger galaxy , the milky way again has a very typical nucleus . ground - based spectroscopic surveys indicate that the majority of galaxies host emission - line nuclei , and the sunns results provide new insights into the structure and energetics of these nebular sources . the findings presented here indicate that the emitting gas is spatially distributed on scales of @xmath28 pc , with accretion - powered objects showing a significant degree of concentration in the observed emission . within these regions , modest radial gradients in optical emission - line ratios are seen ; density gradients evidently play an important role in driving these trends , although there is some evidence as well for ionization gradients in the plasma . liner/ transition nuclei show only mixed indications of a simple composite spatial structure , contrary to expectations , suggesting that distributed sources of ionization beyond normal o - type stars may be involved in powering these common objects . these findings provide a framework for gauging the extent to which the center of the milky way is a typical galaxy nucleus . the galactic center is very similar to other emission - line nuclei in terms of its nebular luminosity , surface brightness distribution , and ionization . combining the present results with other published findings , it is additionally clear that the galactic center resembles the nuclei of other systems in terms of its metallicity and stellar population , and likewise its black hole mass and probable spectroscopic class , considered in relation to the bulge properties of the host galaxy . as a laboratory for the study of galaxy nuclei , the center of the milky way thus provides a very representative example . we thank the referee for constructive comments . support for this research was provided by nasa through grant go-7361 from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . , under nasa contract nas5 - 26555 . cccc ngc 278 & 11.8 & sab(rs)b & h + ngc 2787 & 13.0 & sb(r)0 + & l1.9 + ngc 3351 & 8.1 & sb(r)b & h + ngc 3368 & 8.1 & sab(rs)ab & l2 + ngc 3489 & 6.4 & sab(rs)0 + & t2/s2 + ngc 3982 & 17.0 & sab(r)b : & s1.9 + ngc 3992 & 17.0 & sb(rs)bc & t2 : + ngc 4138 & 17.0 & sa(r)0 + & s1.9 + ngc 4143 & 17.0 & sab(s)0 & l1.9 + ngc 4203 & 9.7 & sab0- : & l1.9 + ngc 4245 & 9.7 & sb(r)0/a : & h + ngc 4314 & 9.7 & sb(rs)a & l2 + ngc 4321 & 16.8 & sab(s)bc & t2 + ngc 4380 & 16.8 & sa(rs)b ? & h + ngc 4435 & 16.8 & sb(s)0 & t2/h : + ngc 4450 & 16.8 & sa(s)ab & l1.9 + ngc 4459 & 16.8 & sa(r)0 + & t2 : + ngc 4477 & 16.8 & sb(s)0 ? & s2 + ngc 4501 & 16.8 & sa(rs)b & s2 + ngc 4548 & 16.8 & sb(rs)b & l2 + ngc 4596 & 16.8 & sb(r)0 + & l2 : : + ngc 4698 & 16.8 & sa(s)ab & s2 + ngc 4800 & 15.2 & sa(rs)b & h + ngc 5055 & 7.2 & sa(rs)bc & t2 + ngc 278 & @xmath95 & @xmath95 & @xmath95 & @xmath95 & @xmath96 & @xmath96 & @xmath96 & @xmath96 & @xmath96 & @xmath97 + ngc 2787 & 390@xmath9840 & 106@xmath9811 & 200@xmath9820 & 280@xmath9830 & 118@xmath9812 & 180@xmath9820 & 500@xmath9850 & 200@xmath9820 & 210@xmath9820 & [email protected] + ngc 3351 & @xmath99 & @xmath99 & 22@xmath983 & 25@xmath983 & @xmath100 & 126@xmath9813 & 71@xmath987 & @xmath101 & @xmath101 & 19@xmath983 + ngc 3368 & 67@xmath9811 & @xmath102 & @xmath102 & @xmath102 & @xmath103 & 117@xmath9812 & 139@xmath9814 & 59@xmath9811 & 59@xmath9811 & [email protected] + ngc 3489 & @xmath104 & @xmath104 & @xmath104 & @xmath104 & @xmath105 & @xmath105 & @xmath105 & @xmath105 & @xmath105 & @xmath106 + ngc 3982 & 560@xmath9860 & 130@xmath9815 & 240@xmath9840 & 5000@xmath98500 & 620@xmath9860 & 1400@xmath98140 & 1380@xmath98140 & 310@xmath9830 & 460@xmath9850 & 180@xmath9820 + ngc 3992 & 61@xmath986 & 12@xmath984 & @xmath107 & 36@xmath984 & @xmath101 & 44@xmath986 & 87@xmath989 & 37@xmath986 & 32@xmath986 & [email protected] + ngc 4143 & 690@xmath9870 & @xmath108 & 400@xmath9840 & 900@xmath9890 & 750@xmath9870 & 390@xmath9840 & 1070@xmath98110 & 340@xmath9830 & 340@xmath9830 & 11@xmath982 + ngc 4203 & 950@xmath98100 & 450@xmath9840 & 900@xmath9890 & 1700@xmath98170 & 1070@xmath98110 & 450@xmath9840 & 1080@xmath98110 & 250@xmath9830 & 340@xmath9830 & 10@xmath981 + ngc 4245 & @xmath109 & @xmath109 & @xmath109 & @xmath109 & @xmath110 & @xmath110 & @xmath110 & @xmath110 & @xmath110 & @xmath111 + ngc 4314 & @xmath112 & @xmath112 & @xmath112 & @xmath112 & @xmath113 & 32@xmath987 & @xmath113 & @xmath113 & @xmath113 & [email protected] + ngc 4321 & 100@xmath9810 & 53@xmath986 & 140@xmath9814 & 47@xmath986 & 106@xmath9813 & 740@xmath9870 & 700@xmath9870 & 112@xmath9811 & 132@xmath9813 & 37@xmath985 + ngc 4380 & @xmath99 & @xmath99 & @xmath99 & @xmath99 & @xmath109 & @xmath109 & @xmath109 & @xmath109 & @xmath109 & @xmath114 + ngc 4435 & @xmath115 & @xmath115 & @xmath115 & @xmath115 & @xmath113 & 28@xmath987 & 63@xmath987 & @xmath113 & @xmath113 & [email protected] + ngc 4450 & 800@xmath9880 & 220@xmath9820 & 400@xmath9840 & 800@xmath9880 & 500@xmath9850 & 400@xmath9840 & 770@xmath9880 & 420@xmath9840 & 460@xmath9850 & 12@xmath982 + ngc 4459 & 42@xmath987 & @xmath113 & @xmath113 & 27@xmath987 & @xmath116 & 63@xmath9812 & 250@xmath9820 & @xmath116 & 38@xmath9812 & [email protected] + ngc 4477 & 86@xmath989 & 14@xmath984 & 28@xmath984 & 82@xmath988 & 26@xmath989 & 110@xmath9811 & 240@xmath9820 & 62@xmath989 & 62@xmath989 & [email protected] + ngc 4501 & 98@xmath9810 & @xmath115 & 36@xmath985 & 163@xmath9816 & 54@xmath989 & 152@xmath9815 & 360@xmath9840 & 96@xmath9810 & 92@xmath989 & [email protected] + ngc 4548 & 55@xmath986 & @xmath117 & @xmath117 & 37@xmath984 & 23@xmath987 & 48@xmath987 & 110@xmath9811 & 48@xmath987 & 40@xmath987 & [email protected] + ngc 4596 & @xmath109 & @xmath109 & @xmath109 & @xmath109 & @xmath95 & 25@xmath988 & 69@xmath988 & @xmath95 & @xmath95 & @xmath118 + ngc 4698 & 90@xmath989 & 40@xmath985 & 60@xmath986 & 149@xmath9815 & @xmath104 & 150@xmath9815 & 200@xmath9820 & 60@xmath986 & 70@xmath987 & [email protected] + ngc 4800 & @xmath112 & @xmath112 & @xmath112 & @xmath112 & @xmath119 & @xmath119 & @xmath119 & @xmath119 & @xmath119 & @xmath97 + ngc 5055 & @xmath120 & @xmath120 & @xmath120 & @xmath120 & @xmath121 & @xmath121 & @xmath121 & @xmath121 & @xmath121 & @xmath106 + ccccc ngc 2787 & @xmath122 & 0.35 & @xmath123 & @xmath124 + ngc 3982 & @xmath125 & 0.12 & @xmath126 & @xmath127 + ngc 4143 & @xmath128 & 0.45 & @xmath129 & @xmath130 + ngc 4203 & @xmath131 & 0.34 & @xmath132 ( @xmath133 ) & @xmath134 ( 0.89 ) + ngc 4450 & ... & 0.20 & @xmath135 ( @xmath136 ) & @xmath137 ( 0.82 ) + ngc 4501 & @xmath138 & 0.09 & @xmath139 & @xmath140 +

we present results from a program of optical spectroscopy for 23 nearby galaxies with emission - line nuclei . this investigation takes advantage of the spatial resolution of the _ hubble space telescope _ to study the structure and energetics of the central @xmath0 pc , and the resulting data have value for quantifying central black hole masses , star formation histories , and nebular properties . this paper provides a description of the experimental design , and new findings from the study of emission lines . the sample targets span a range of nebular spectroscopic class , from to seyfert nuclei . this data set and the resulting measurements are unique in terms of the sample size , the range of nebular class , and the investigation of physical scales extending down to parsecs . the line ratios indicative of nebular ionization show only modest variations over order - of - magnitude differences in radius , and demonstrate in a systematic way that geometrical dilution of the radiation field from a central source can not be assumed as a primary driver of ionization structure . comparisons between large- and small - aperture measurements for the /liner transition objects provide a new test that challenges conventional wisdom concerning the composite nature of these systems . we also list a number of other quantitative results that are of interest for understanding galaxy nuclei , including ( 1 ) the spatial distribution / degree of concentration of h@xmath1 emission as a function of nebular type ; ( 2 ) the radial variation in electron density as a function of nebular type ; and ( 3 ) quantitative broad h@xmath1 estimates obtained at a second epoch for these low - luminosity nuclei . the resulting measurements provide a new basis for comparing the nuclei of other galaxies with that of the milky way . we find that the galactic center is representative across a wide span of properties as a low - luminosity emission - line nucleus .

the central point of these lectures is the availability of a field theoretical framework for the strong interactions and its use to study the structure of hadrons . controlling and selectively probing the nonperturbative regime in high energy scattering processes is the key to study the structure of hadrons in the context of qcd . the control parameters for the target and the probe are the spin and flavor , which in combination with the kinematical flexibility in scattering processes is used to select the observable and its gluonic or quarkic nature . examples are aaaaaaaaaaaaaaaa = bbbbbbbbbbbbbbbbbbbbbb = cccccccccccc @xmath0 ( elastic leptoproduction ) ( spacelike ) form factors , + @xmath1 ( inclusive leptoproduction ) distribution functions , + @xmath2 ( 1-particle inclusive distribution and + leptoproduction ) fragmentation functions , + @xmath3 ( annihilation into @xmath4 ) ( timelike ) form factors , + @xmath5 ( 1-particle inclusive fragmentation functions , + annihilation ) + @xmath6 ( drell - yan scattering ) distribution functions . these notes focus on leptoproduction in which one deals both with distribution functions ( in inclusive processes ) and fragmentation functions ( in semi - inclusive processes ) . for the process @xmath7 ( see fig . [ fig0 ] ) , the cross section can be separated into a lepton and hadron part . although the lepton part is simpler , let us start with the hadron part , & & 2mw_^(h ) ( q ; p s ) = _ x ( 2)^4 ^4 ( q + p - p_x ) + & & |j_(0)|p_x p_x |j_(0)|p s , @xmath8 @xmath9 the simplest thing to do is to parametrize this tensor in standard tensors and structure functions . instead of the traditional choice @xcite for these tensors , @xmath10 and @xmath11 and structure functions @xmath12 and @xmath13 , we immediately go to a dimensionless representation , using a natural space - like momentum ( defined by @xmath14 ) and a time - like momentum constructed from @xmath15 and @xmath14 , & & z^= -q^= - , + & & t^= = ( p^- q^ ) = . using hermiticity for the currents , parity invariance and current conservation one obtains as the most general form the symmetric tensor mw_s^(q , p ) = _ -g_^f_1 + t^t^_f_l , [ param ] where the structure functions @xmath16 and @xmath17 or the transverse and longitudinal structure functions , @xmath18 and @xmath19 , depend only on the for the hadron part relevant invariants @xmath20 and @xmath21 . in all equations given here we have omitted target mass effects of order @xmath22 . in order to write down the cross section one needs to include the necessary phase space factors and include the lepton part given by the tensor l_^(h ) ( k ; k^^ ) = 2 k_k^ _ + 2 k_k^ _ - q^2 g _ + 2i_e _ q^k^ . we have included here the ( longitudinal ) lepton polarization ( @xmath23 ) . for later convenience it is useful to rewrite this tensor also in terms of the space - like and time - like vectors @xmath24 and @xmath25 . it is a straightforward exercise to get k^= q^+ t^+ ^ , where @xmath26 is the perpendicular direction defining the lepton scattering plane ( see fig . [ fig1 ] ) . this perpendicular direction becomes relevant only if other vectors than @xmath15 and @xmath14 are present , e.g. a spin direction in polarized scattering or the momentum of a produced hadron in 1-particle inclusive processes . the lepton tensor becomes l^_(h ) & = & , [ leptontensor ] where @xmath27 . the cross section for unpolarized lepton and hadron only involves the first two ( symmetric ) terms in the lepton tensor and one obtains & = & \ { ( 1-y + y^2 ) f_t(x__b , q^2 ) + ( 1 -y ) f_l(x__b , q^2)}. as soon as the exchange of a @xmath28 boson becomes important the hadron tensor is no longer constrained by parity invariance and a third structure function @xmath29 becomes important . the use of polarization in leptoproduction provides new ways to probe the hadron target . for a spin 1/2 particle the initial state is described by a 2-dimensional spin density matrix @xmath30 describing the probabilities @xmath31 for a variety of spin possibilities . this density matrix is hermitean with tr@xmath32 = 1 . it can in the target rest frame be expanded _ ss^ = ( 1 + s_ss^ ) , where @xmath33 is the spin vector . when @xmath34 one has a pure state ( only one state @xmath35 and @xmath36 ) , when @xmath37 one has an ensemble of states . for the case @xmath38 one has simply an averaging over spins , corresponding to an unpolarized ensemble . to include spin one could generalize the hadron tensor to a matrix in spin space , @xmath39 @xmath40 @xmath41 depending only on the momenta or one can look at the tensor @xmath42 , which is given by w^(q , p , s ) = ( ( p , s ) w^(q , p ) ) , [ defs ] with the spacelike spin vector @xmath43 appearing _ linearly _ and in an arbitrary frame satisfying @xmath44 . it has invariant length @xmath45 . it is convenient to write the spin vector as s^= ( p^- q^ ) + s_^ , [ inclspin ] with @xmath46 . for a pure state one has @xmath47 . using symmetry constraints one obtains for electromagnetic interactions ( parity conservation ) an antisymmetric part in the hadron tensor , mw_a^(q , p , s ) = _ -i_^ g_1 + i t_^[_^ ] s _ g_t . [ wanti ] the polarized part of the cross section becomes & = & _ e \ { ( 1- ) g_1(x__b , q^2 ) + & & -_s^ g_t(x__b , q^2)}. more flexibility in probing new aspects of hadron structure is achieved in semi - inclusive scattering processes . for instance in 1-particle inclusive measurements one can measure azimuthal dependences in the cross sections . the central object of interest for 1-particle inclusive leptoproduction , the hadron tensor , is given by & & 2mw_^(h ) ( q ; p s ; p_h s_h ) = _ x ^4 ( q + p - p_x - p_h ) + & & |j_(0)|p_x ; p_h s_h p_x ; p_h s_h |j_(0)|p s , where @xmath48 and @xmath49 are the momenta and spin vectors of target hadron and produced hadron , @xmath14 is the ( space - like ) momentum transfer with @xmath50 = @xmath20 sufficiently large . the kinematics is illustrated in fig . [ fig1 ] , where also the scaling variables are introduced . @xmath51 @xmath52 for the parametrization of the hadron tensor in terms of structure functions it is useful to introduce the directions @xmath24 and @xmath25 as before and using the vector @xmath53 to construct a vector that is orthogonal to these vectors . for the situation that @xmath54 is @xmath55 ( current fragmentation ! ) one finds that q_^= q^+x__bp^- = - -q_th^ , is such a vector . this vector is proportional to the transverse momentum of the outgoing hadron with respect ot @xmath15 and @xmath14 . it can also be considered as the transverse momentum of the photon with respect to the hadron momenta @xmath15 and @xmath53 . for an unpolarized ( or spin 0 ) hadron in the final state the symmetric part of the tensor is given by mw_s^(q , p , p_h ) & = & - g_^h_t + t^t^h_l + & & + t^\{h^}h_lt + 2h^h^+ g_^_tt . noteworthy is that also an antisymmetric term in the tensor is allowed , mw_a^(q , p , p_h ) = - it^[h^]h^_lt . [ sidiswanti ] clearly the lepton tensor in eq . [ leptontensor ] is able to distinguish all the structures in the semi - inclusive hadron tensor . the symmetric part gives the cross section for unpolarized leptons , & = & x__bz_h\ { ( 1-y+y^2 ) h_t + ( 1-y)h_l + & & - ( 2-y)_h^h_lt + & & + ( 1-y)2_h^h_tt } while the antisymmetric part gives the cross section for a polarized lepton ( note the target is not polarized ! ) = _ ez_h_h^h^_lt . of course many more structure functions appear for polarized targets or if one considers polarimetry in the final state . within the framework of qcd and knowing that the photon or @xmath28 current couples to the quarks , it is possible to write down a diagrammatic expansion for leptoproduction , with in the deep inelastic limit ( @xmath56 ) as relevant diagrams only the ones given in fig . [ fig2 ] for inclusive and 1-particle inclusive scattering respectively . the expression for @xmath57 can be rewritten as a nonlocal product of currents and it is a straightforward exercise to show by inserting the currents @xmath58 that for 1-particle inclusive scattering one obtains in tree approximation & & 2mw _ ( q ; p s ; p_h s_h ) = + & & = d^4x e^iqx p s |:_j ( x ) ( _ ) _ jk _ k(x ) : _ x |x ; p_h s_h + & & x ; p_h s_h | : _ l(0 ) ( _ ) _ li _ i(0):|p s + & & = d^4x e^iqx p s |_j ( x ) _ i(0 ) p s ( _ ) _ jk + & & 0 _ k(x ) _ x |x ; p_h s_h x ; p_h s_h | _ l(0 ) 0 ( _ ) _ li + & & + d^4x e^iqx p s |_k ( x ) _ l(0 ) p s ( _ ) _ li + & & 0 _ j(x ) _ x |x ; p_h s_h x ; p_h s_h | _ i(0 ) 0 ( _ ) _ jk , + & & = d^4p d^4k^4(p+q - k ) ( ( p ) _ ( k ) _ ) + \ { c q -q + } , [ basic ] where & & _ ij(p ) = d^4 e^ip p s |_j ( 0 ) _ i()|p s , + & & _ kl(k ) = d^4 e^ik 0 _ k ( ) _ x |x ; p_h s_h x ; p_h s_h | _ l(0 ) 0 . note that in @xmath59 ( quark production ) a summation over colors is assumed , while in @xmath60 ( quark decay ) an averaging over colors is assumed . the quantities @xmath59 and @xmath60 correspond to the blobs in fig . [ fig2 ] and parametrize the soft physics . soft refers to all invariants of momenta being small as compared to the hard scale , i.e. for @xmath61 one has @xmath62 . in general many more diagrams have to be considered in evaluating the hadron tensors , but in the deep inelastic limit they can be neglected or considered as corrections to the soft blobs . we return to this later . as mentioned above , the relevant structural information for the hadrons is contained in soft parts ( the blobs in fig . [ fig2 ] ) which represent specific matrix elements of quark fields . the form of @xmath59 is constrained by hermiticity , parity and time - reversal invariance . the quantity depends besides the quark momentum @xmath63 on the target momentum @xmath15 and the spin vector @xmath43 and one must have @xmath64 } \quad \rightarrow \quad \phi^\dagger ( p , p , s ) = \gamma_0 \,\phi(p , p , s)\,\gamma_0 , \\ & & \mbox{[parity ] } \quad \rightarrow \quad \phi(p , p , s ) = \gamma_0 \,\phi(\bar p,\bar p,-\bar s)\,\gamma_0 , \\ & & \mbox{[time\ reversal ] } \ \rightarrow \ \phi^\ast(p , p , s ) = ( -i\gamma_5 c)\,\phi(\bar p,\bar p , \bar s)\,(-i\gamma_5 c ) , \end{aligned}\ ] ] where @xmath65 = @xmath66 , @xmath67= @xmath68 and @xmath69 = @xmath70 . the most general way to parametrize @xmath59 using only the constraints from hermiticity and parity invariance , is @xcite ( p , p , s ) & = & ma_1 + a_2/p + a_3 /p + ia_4 + & & + ia_5(ps ) _ 5 + ma_6 /s _ 5 + a_7/p _ 5 + & & + a_8/p _ 5 + a_9_5 + a_10_5 + & & + a_11_5 + a_12 , [ lorentz ] where the first four terms do not involve the hadron polarization vector . hermiticity requires all the amplitudes @xmath71 = @xmath72 to be real . the amplitudes @xmath73 , @xmath74 and @xmath75 vanish when also time reversal invariance applies . in order to find out which information in the soft parts is important in a hard process one needs to realize that the hard scale @xmath76 leads in a natural way to the use of lightlike vectors @xmath77 and @xmath78 satisfying @xmath79 and @xmath80 = 1 . for inclusive scattering one parametrizes the momenta @xmath81 the above are the external momenta . next turn to the internal momenta , looking at the left diagram in fig . [ fig2 ] . in the soft part actually _ all _ momenta , that is @xmath63 _ and _ @xmath15 have a minus component that can be neglected compared to that in the hard part , since otherwise @xmath82 would be hard . thus because @xmath63 must have only a hard plus component , @xmath14 has two hard components and @xmath83 being the current jet also must be soft , i.e. only can have one large lightcone component , one must have @xmath84 where the parts indicate ( negligible ) @xmath85 terms . also the transverse component is not relevant for the hard part . one thus sees that for inclusive scattering the only relevant dependence of the soft part is the @xmath86 dependence . moreover , the above requirements on the internal momenta already indicate that the lightcone fraction @xmath87 must be equal to @xmath21 . this will come out when we do the actual calculation in one of the next sections . the minus component @xmath88 and transverse components thus can be integrated over restricting the nonlocality in @xmath61 . the relevant soft part then is some dirac trace of the quantity @xcite _ ij(x ) & = & dp^-d^2p _ _ ij(p , p , s ) + & = & . e^ip p , s_j(0 ) _ i ( ) p , s|_^+ = _ = 0 , depending on the lightcone fraction @xmath87 . to be precise one puts in the full form for the quark momentum , p = xp^+n_+ + n_- + p _ , [ quarkmom ] and performs the integration over @xmath61 using dp^-d^2p_ = d(pp)dp^2 . when one wants to calculate the leading order in @xmath85 for a hard process , one only needs to look at leading parts in @xmath89 because @xmath90 ( see opening paragraph of this section ) @xcite . in this case that turns out to be the part proportional to @xmath91 , ( x ) = \ { f_1(x)0.2 em n-0.4em /_+ + g_1(x ) _ 50.2 em n-0.4em /_+ + h_1(x ) } + o ( ) the precise expression of the functions @xmath92 , etc . as integrals over the amplitudes can be easily written down . using field theoretical methods the left diagram in fig . [ fig2 ] can now be calculated . omitting the sum over flavors ( @xmath93 ) , the quark charges @xmath94 and the @xmath95 antiquark diagram , the symmetric part of the hadron tensor the result is 2mw^(p , q ) = dp^-dp^+d^2p _ ( ( p ) ^(p+q ) ^ ) , where ( k ) = ( /k + m)(k^2-m^2 ) ( k^+ ) , and in the approximation anything proportional to @xmath96 has been neglected . one obtains @xmath97 antiquarks arise from the diagram with opposite fermion flow , proportional to @xmath98 with _ ij(p ) = d^4 e^-ip p s |_i ( ) _ j(0)|p s . the _ proper _ definition of antiquark distributions starts from @xmath99 containing antiquark distributions @xmath100 , etc . the quantity @xmath101 is obtained from @xmath61 after the replacement of @xmath102 by @xmath103 . one then finds @xmath104 = @xmath105 , i.e. one has to be aware of sign differences . symmetry relations between quark and antiquark relations can be obtained using the anticommutation relations for fermions , giving @xmath106 . one finds that @xmath107 , @xmath108 , and @xmath109 . finally , after including the flavor summation and the quark charges squared one can compare the result with eq . [ param ] to obtain for the structure function 2f_1(x__b ) = _ a e_a^2(f_1^a(x__b ) + f_1^|a(x__b ) ) , while @xmath110 ( callan - gross relation ) . the antisymmetric part of @xmath111 in the above calculation is left as an exercise . the answer is @xmath112 which after inclusion of antiquarks , flavor summation gives ( cf eq . [ wanti ] ) 2g_1(x__b ) = _ a e_a^2(g_1^a(x__b ) + g_1^|a(x__b ) ) . the functions @xmath113 , @xmath114 and @xmath115 can be obtained from the correlator @xmath116 after tracing with the appropriate dirac matrix , f_1(x ) & = & . e^ip p , s(0 ) ^+ ( ) p , s|_^+ = _ = 0 , + g_1(x ) & = & . e^ip p , s(0 ) ^+_5 ( ) p , s|_^+ = _ = 0 , + s_^ih_1(x ) & = & . e^ip p , s(0 ) i^i+_5 ( ) p , s|_^+ = _ = 0 , by introducing _ good _ and _ bad _ fields @xmath117 , one sees that @xmath113 can be rewritten as f_1(x ) & = & . e^ip p , s^_+(0 ) _ + ( ) p , s|_^+ = _ = 0 + & = & _ n | p_n_+p|^2(p_n^+ - ( 1-x)p^+ ) , i.e. it is a quark lightcone momentum distribution . for the functions @xmath114 and @xmath115 one needs in addition the projectors on quark chirality states , @xmath118 , and on quark transverse spin states @xcite , @xmath119 to see that & & f_1(x ) = f_1r(x ) + f_1l(x ) = f_1(x ) + f_1(x ) , + & & g_1(x ) = f_1r(x ) - f_1l(x ) , + & & h_1(x ) = f_1(x ) - f_1(x ) . one sees some trivial bounds such as @xmath120 and @xmath121 . since @xmath122 and sees @xmath123 . from the antiquark distribution @xmath100 and its relation to @xmath92 one obtains @xmath124 , thus the support of the functions is @xmath125 . the trivial bounds on the distribution functions ( @xmath126 and @xmath127 ) can be sharpened . for instance one can look explicitly at the structure in dirac space of the correlation function @xmath128 . actually , we will look at the correlation functions @xmath129 , which involves at leading order matrix elements @xmath130 . one has in weyl representation ( @xmath131 , @xmath132 , @xmath133 ) the matrices @xmath134 the good projector only leaves two ( independent ) dirac spinors , one righthanded ( r ) , one lefthanded ( l ) . on this basis of good r and l spinors the for hard scattering processes relevant matrix @xmath135 is given by ( /n_-)_ij(x ) = cc f_1 + g_1 & ( s_^1-is_^2)h_1 + & + ( s_^1+is_^2)h_1 & f_1 - g_1 one can also turn the @xmath43-dependent correlation function @xmath59 defined in analogy with @xmath136 in eq . [ defs ] into a matrix in the nucleon spin space . if ( x ; p , s ) & = & _ o + _ l + s_^1_^1 + s_^2_^2 , then one has on the basis of spin 1/2 target states with @xmath137 and @xmath138 respectively _ ss^(x ) = cc _ o + _ l & _ ^1 - i_^2 + & + _ ^1 + i_^2 & _ o - _ l + the matrix relevant for bounds is the matrix @xmath139 = @xmath140 ( for this matrix one has @xmath141 for any direction @xmath142 ) . on the basis @xmath143 , @xmath144 , @xmath145 and @xmath146 it becomes ( ( x)/n_-)^t & = & cccc f_1 + g_1 & 0 & 0 & 2h_1 + & & & + 0 & f_1 - g_1 & 0 & 0 + & & & + 0 & 0 & f_1 - g_1 & 0 + & & & + 2h_1 & 0 & 0 & f_1 + g_1 c + + + [ prod1 ] + & & of this matrix any diagonal matrix element must always be positive , hence the eigenvalues must be positive , which gives a bound on the distribution functions stronger than the trivial bounds , namely h_1(x ) ( f_1(x ) + g_1(x ) ) known as the soffer bound @xcite . for the functions appearing in the soft parts , and thus also for the structure functions , one can derive sum rules . starting with the traces defining the quark distributions , @xmath147 and integrating over @xmath87 one obtains ( using symmetry relation as indicated above to eliminate antiquarks @xmath148 ) , _ 0 ^ 1 dx ( f_1(x ) - |f_1(x ) ) = _ -1 ^ 1 dx f_1(x ) = , which as we have seen in the section on elastic scattering is nothing else than a form factor at zero momentum transfer , i.e. the number of quarks of that particular flavor . similarly one finds the sum rule _ 0 ^ 1 dx ( g_1(x ) + |g_1(x ) ) = _ -1 ^ 1 dx g_1(x ) = , [ axsumr ] which precisely is the axial charge @xmath149 for a particular quark flavor . these relations of quark distributions and matrix elements underly sum rules for the structure functions , such as the bjorken sum rule _ 0 ^ 1 dx__b ( g_1^p(x__b , q^2)-g_1^n(x__b , q^2 ) ) = ( g_a^u - g_a^d ) = g_a^np(0 ) , relating the polarized structure function to the axial charge measured in neutron @xmath150-decay which also can be expressed in quark axial charges . for 1-particle inclusive scattering one parametrizes the momenta @xmath151 note that this works for socalled current fragmentation , in which case the produced hadron is _ hard _ with respect to the target momentum , i.e. @xmath152 . the minus component @xmath153 is irrelevant in the lower soft part , while the plus component @xmath154 is irrelevant in the upper soft part . note that after the choice of @xmath15 and @xmath53 one can no longer omit a transverse component in the other vector , in this case the momentum transfer @xmath14 . this is precisely the vector @xmath155 introduced earlier in the discussion of the structure functions for 1-particle inclusive leptoproduction . one immediately sees that one can no longer simply integrate over the transverse component of the quark momentum , defined in eq . [ quarkmom ] . at this point it turns out that the most convenient way to describe the spin vector of the target is via an expansion of the form s^= -n_- + n_+ + s_. one has up to @xmath156 corrections @xmath157 and @xmath158 . for a pure state one has @xmath159 , in general this quantity being less or equal than one . the soft part to look at is ( x , p_t ) = . e^ip p , s(0 ) ( ) p , s|_^+ = 0 . for the leading order results , it is parametrized as ( x , p _ ) = _ o(x , p _ ) + _ l(x , p _ ) + _ t(x , p _ ) , with the parts involving unpolarized targets ( o ) , longitudinally polarized targets ( l ) and transversely polarized targets ( t ) up to parts proportional to @xmath89 given by _ o(x , p _ ) & = & \ { f_1(x , p_)0.2 em n-0.4em /_+ + h_1^(x , p _ ) } + _ l(x , p _ ) & = & \ { g_1l(x , p_)_50.2 em n-0.4em /_+ + h_1l^(x , p _ ) } + _ t(x , p _ ) & = & \ { f_1t^(x , p _ ) + & & + g_1t(x , p _ ) _ 50.2 em n-0.4em /_+ + h_1t(x , p _ ) + & & + h_1t^(x , p _ ) } . all functions appearing here have a natural interpretation as densities . this is seen as discussed before for the @xmath160-integrated functions . now it includes densities such as the density of longitudinally polarized quarks in a transversely polarized nucleon ( @xmath161 ) and the density of transversely polarized quarks in a longitudinally polarized nucleon ( @xmath162 ) . the interpretation of all functions is illustrated in fig . [ fig3 ] . several functions vanish from the soft part upon integration over @xmath163 . actually we will find that particularly interesting functions survive when one integrates over @xmath160 weighting with @xmath164 , e.g. _ ^(x ) d^2 p _ ( x , p _ ) & = & \ { -g_1t^(1)(x)s_^/n_+_5 -h_1l^(1)(x ) + & & -f_1t^(1 ) ^ _ ^n_-^ - h_1^(1 ) } , [ phid ] where we define @xmath165-moments , referred to as transverse moments , as g_1t^(n)(x ) = d^2p _ ( ) ^n g_1t(x , p _ ) , and similarly the other functions . the use of these moments also turns out to be important when one tries to incorporate perturbative effects that affect the @xmath163-shape @xcite . to interpret all functions it is simplest to again write down the matrix @xmath166 , which becomes ( /n_-)^t = cccc f_1 + g_1 & e^ig_1 t & e^-ih_1l^ & 2h_1 + & & & + e^ig_1t^ & f_1 - g_1 & e^-2ih_1t^ & -e^-ih_1l^ + & & & + e^ih_1l^ & e^2ih_1t^ & f_1 - g_1 & -e^ig_1t^ + & & & + 2h_1 & -e^ih_1l^ & -e^-ig_1 t & f_1 + g_1 , to be compared with eq . [ prod1 ] . we have omitted here the t - odd functions @xmath167 and @xmath168 appearing as imaginary parts of @xmath169 and @xmath162 , respectively . the functions @xmath168 and @xmath167 are t - odd , vanishing if t - reversal invariance can be applied to the matrix element . for @xmath163-dependent correlation functions , matrix elements involving gluon fields at infinity ( gluonic poles @xcite ) can for instance prevent application of t - reversal invariance . the functions describe the possible appearance of unpolarized quarks in a transversely polarized nucleon ( @xmath167 ) or transversely polarized quarks in an unpolarized hadron ( @xmath168 ) and lead to single - spin asymmetries in various processes @xcite . the interpretation of all these functions is also illustrated in fig . [ fig3 ] . + + + of course just integrating @xmath170 over @xmath163 gives the result used in inclusive scattering with @xmath92 = @xmath171 , @xmath172 = @xmath173 and @xmath174 = @xmath175 . we note that the function @xmath176 appears after weighting with @xmath177 . in analogy to the soffer bound derived from the production matrix in eq . [ prod1 ] one easily derives a number of new bounds from the full matrix , such as & & f_1(x , p_^2 ) 0 , + & & g_1(x , p_^2)f_1(x , p_^2 ) . obtained from one - dimensional subspaces and & & h_1 ( f_1 + g_1 ) f_1 , [ soffer ] + & & h_1t^(1 ) ( f_1 - g_1 ) f_1 , + & & g_1t^(1)^2 ( f_1 + g_1 ) ( f_1 - g_1 ) f_1 ^ 2 , + & & h_1l^(1)^2 ( f_1 + g_1 ) ( f_1 - g_1 ) f_1 ^ 2 , obtained from two - dimensional subspaces . here we have used the notation @xmath178 also for @xmath163-dependent functions . these bounds and their further refinements have been discussed in detail in ref . there are straightforward extensions of transverse momentum dependent distribution and fragmentation functions for spin 1 hadrons @xcite and gluons in spin 1/2 hadrons @xcite . just as for the distribution functions one can perform an analysis of the soft part describing the quark fragmentation . one needs @xcite _ ij(z , k ) = . _ x e^ik tr 0 _ i ( ) p_h , xp_h , x_j(0 ) 0 |_^- = 0 . for the production of unpolarized hadrons @xmath179 in hard processes one needs to leading order in @xmath85 the correlation function , _ o(z , k _ ) = zd_1(z , k^_)0.2 em n-0.4em /_- + zh_1^(z , k^ _ ) + o ( ) . when we limit ourselves to an unpolarized or spin 0 final state hadron . the arguments of the fragmentation functions @xmath180 and @xmath181 are @xmath182 = @xmath183 and @xmath184 = @xmath185 . the first is the ( lightcone ) momentum fraction of the produced hadron , the second is the transverse momentum of the produced hadron with respect to the quark . the fragmentation function @xmath180 is the equivalent of the distribution function @xmath113 . it can be interpreted as the probability of finding a hadron @xmath179 in a quark . the function @xmath181 , interpretable as the difference in production probabilities of unpolarized hadrons from a transversely polarized quark depending on transverse momentum , is allowed because of the non - applicability of time reversal invariance @xcite . this is natural for the fragmentation functions @xcite because of the appearance of out - states @xmath186 in the definition of @xmath60 , in contrast to the plane wave states appearing in @xmath59 . after @xmath187-averaging one is left with the functions @xmath188 and the @xmath189-weighted result @xmath190 . after the analysis of the soft parts , the next step is to find out how one obtains the information on the various correlation functions from experiments , in this particular case in lepton - hadron scattering via one - photon exchange as discussed before . to get the leading order result for semi - inclusive scattering it is sufficient to compute the diagram in fig . [ fig2 ] ( right ) by using qcd and qed feynman rules in the hard part and the matrix elements @xmath59 and @xmath60 for the soft parts , parametrized in terms of distribution and fragmentation functions . the most well - known results for leptoproduction are : & & = _ a,|a e_a^2 1 + ( 1-y)^2x__bf^a_1(x__b ) d^a_1(z_h ) + & & = _ e _ e_a^2 y ( 2-y ) x__bg^a_1(x__b)d^a_1(z_h ) the indices attached to the cross section refer to polarization of lepton ( o is unpolarized , l is longitudinally polarized ) and hadron ( o is unpolarized , l is longitudinally polarized , t is transversely polarized ) . note that the result is a weighted sum over quarks and antiquarks involving the charge @xmath191 squared . comparing with well - known formal expansions of the cross section in terms of structure functions one can simply identify these . for instance the above result for unpolarized scattering ( oo ) shows that after averaging over azimuthal angles , only one structure function survives if we work at order @xmath192 and at leading order in @xmath85 . as we have seen , in 1-particle inclusive unpolarized leptoproduction in principle four structure functions appear , two of them containing azimuthal dependence of the form @xmath193 and @xmath194 . the first one only appears at order @xmath85 @xcite , the second one even at leading order but only in the case of the existence of nonvanishing t - odd distribution functions . to be specific if we define weighted cross section such as d^2q_t ( 2_h^ ) _ oo we obtain the following asymmetry , . _ oo = ( 1-y)_a,|a e_a^2 x__bh_1^(1)a(x__b ) h_1^(1)a . in lepton - hadron scattering this asymmetry requires t - odd distribution functions and therefore most likely is absent or very small . in @xmath195 annihilation @xcite , however , a @xmath196 asymmetry between produced particles ( e.g. pions ) in opposite jets involves two very likely nonvanishing fragmentation functions @xmath181 and @xmath197 . indications for the presence of these fragmentation functions have been found in lep data @xcite . also in leptoproduction indications for azimuthal asymmetries have been found @xcite , which in case of single spin asymmetries point towards a t - odd fragmentation function . for polarized targets , several azimuthal asymmetries arise already at leading order . for example the following possibilities were investigated in refs @xcite . & & _ lt = _ es y(2-y)_a,|a e_a^2 x__bg_1t^(1)a(x__b ) d^a_1(z_h ) , + & & [ asbas ] + & & _ ol = - ( 1-y)_a,|a e_a^2 x__bh_1l^(1)a(x__b ) h_1^(1)a(z_h ) , + & & [ as2 ] + & & _ ot = s ( 1-y)_a,|a e_a^2 x__bh_1^a(x__b ) h_1^(1)a(z_h ) . + & & [ finalstate ] the latter two are single spin asymmetries involving the fragmentation function @xmath198 . the last one was the asymmetry proposed by collins @xcite as a way to access the transverse spin distribution function @xmath115 in pion production . note , however , that in using the azimuthal dependence one needs to be very careful . for instance , besides the @xmath199 , one also finds at leading order a @xmath200 asymmetry which is proportional to @xmath201 @xcite . if one proceeds up to order @xmath85 one also needs terms in the parametrization of the soft part proportional to @xmath89 . limiting ourselves to the @xmath160-integrated correlations one needs @xmath202}{2}\biggr\ } \nonumber \\ & + & \frac{m}{2p^+}\biggl\ { e(x ) + g_t(x)\,\gamma_5\,{\kern 0.2 em s\kern -0.56em \raisebox{0.3ex}{/}}_\st + \lambda\,h_l(x)\,\frac{\gamma_5\,[{\kern 0.2 em n\kern -0.4em /}_+,{\kern 0.2 em n\kern -0.4em /}_-]}{2 } \biggr\ } \nonumber \\ & + & \frac{m}{2p^+}\biggl\ { -\lambda\,e_l(x)\,i\gamma_5 - f_t(x)\,\epsilon_\st^{\rho\sigma}\gamma_\rho s_{\st\sigma } + h(x)\,\frac{i\,[{\kern 0.2 em n\kern -0.4em /}_+,{\kern 0.2 em n\kern -0.4em /}_-]}{2 } \biggr\}.\end{aligned}\ ] ] the last set of three terms proportional to @xmath89 vanish when time - reversal invariance applies . actually in the calculation of the cross section one has to be careful . let us use inclusive scattering off a transversely polarized nucleon ( transverse means @xmath203 in eq . [ inclspin ] ) as an example . the hadronic tensor is zero in leading order in @xmath85 . at order @xmath85 one obtains from the handbag diagram a contribution 2mw_a ( a)^(q , p , s _ ) = i t_^[_^ ] s _ ( g_1t^(1)(x__b ) - h_1(x__b ) ) . it shows that one must be very careful with the integration over @xmath163 . there is a second contribution at order @xmath85 coming from diagrams as the one shown in fig . [ figdis1 ] . for these gluon diagrams one needs matrix elements containing @xmath204 . at order @xmath85 one only needs the matrix element of the bilocal combinations @xmath205 and @xmath206 these soft parts have a structure quite similar to @xmath207 and are parametrized as _ a^(x ) & = & \ { -xg_t(x)s_^/n_+_5 -xh_l(x ) + & & -xf_t(x ) ^ _ ^n_-^ - xh(x ) } . [ phia ] this contributes also to @xmath208 , 2mw_a ( b)^(q , p , s _ ) = i t_^[_^ ] s _ g_t(x__b ) . using the qcd equations of motion , however , these functions can be related to the functions appearing in @xmath59 . to be precise one combines @xmath209 in @xmath210 ( see eq . [ phid ] ) and @xmath211 in @xmath212 to @xmath213 containing @xmath214 for which one has via the equations of motion _ d^(x ) & = & \ { -(xg_t- h_1)s_^/n_+_5 + & & -(xh_l - g_1 ) + & & -xf_t(x ) ^ _ ^n_-^ - xh(x ) } . hence one obtains & & xg_t = xg_t - g_1t^(1 ) - h_1 , + & & xh_l = xh_l - h_1l^(1 ) - g_1 , + & & xf_t = xf_t + f_1t^(1 ) , + & & xh = xh + 2h_1^(1 ) . and one obtains the full contribution 2mw_a^(q , p , s _ ) = i t_^[_^ ] s _ g_t(x__b ) , leading for the structure function @xmath215 defined in eq . [ wanti ] to the result _ t(x__b , q^2 ) = _ a e_a^2 ( g_t^a(x__b ) + g_t^|a(x__b ) ) . from lorentz invariance one obtains , furthermore , some interesting relations between the subleading functions and the @xmath216-dependent leading functions @xcite . just by using the expressions for the functions in terms of the amplitudes @xmath71 in eq . [ lorentz ] one finds & & g_t = g_1 + g_1t^(1 ) , [ gtrel ] + & & h_l = h_1 - h_1l^(1 ) , [ hlrel ] + & & f_t = - f_1t^(1 ) , + & & h = - h_1^(1 ) . [ rel4 ] as an application , one can eliminate @xmath217 using eq . [ gtrel ] and obtain ( assuming sufficient neat behavior of the functions ) for @xmath218 g_2(x ) & = & -+ + & & + . [ wweq ] one can use this to obtain for each quark flavor @xmath219 , the burkhardt - cottingham sumrule @xcite . neglecting the interaction - dependent part one obtains the wandzura - wilczek approximation @xcite for @xmath220 , which in particular when one neglects the quark mass term provides a simple and often used estimate for @xmath220 . it has become the standard with which experimentalists compare the results for @xmath220 . actually the slac results for @xmath220 can also be used to estimate the function @xmath217 and the resulting asymmetries , e.g. the one in eq . [ asbas ] . for this one needs the exact relation in eq . [ gtrel ] . results can be found in refs @xcite and @xcite . the full set of results for the twist-3 functions , including as the first one the above wandzura - wilczek result in eq . [ wweq ] are ( omitting quark mass terms ) & & g_t(x ) = _ x^1 dy + , + & & = _ x^1 dy - _ x^1 dy , + & & h_l(x ) = 2x_x^1 dy + , + & & = -_x^1 dy + _ x^1 dy , + & & f_t(x ) = , + & & = _ x^1 dy , + & & h(x ) = , + & & = _ x^1 dy . the wandzura - wilczek result has a complete analogue in the relation for @xmath221 discussed in ref . @xcite . in slightly different form the result for @xmath217 has been discussed in ref . actually , we need not consider the t - odd functions separately . they can be simply considered as imaginary parts of other functions , when we allow complex functions . in particular one can expand the correlation functions into matrices in dirac space @xcite to show that the relevant combinations are @xmath222 which we can treat together as one complex function @xmath161 . similarly we can use complex functions @xmath223 @xmath224 @xmath162 , @xmath225 @xmath224 @xmath226 , @xmath227 @xmath224 @xmath221 , @xmath228 @xmath224 @xmath229 . the functions @xmath113 , @xmath114 and @xmath115 remain real , they do nt have t - odd partners . also for the transverse momentum dependent functions dependent distribution and fragmentation functions one can proceed to subleading order @xcite , but very likely one will also find competing @xmath230 corrections contributing to the same observables . we will not discuss these functions here . in semi - inclusive cross sections one also needs fragmentation functions , for which similar relations exist , e.g. the relation in eq . [ rel4 ] for distribution functions has an analog for fragmentation functions , relating @xmath198 ( appearing in eqs [ as2 ] and [ finalstate ] ) and an at subleading order appearing function @xmath231 , = z^2 ( ) . [ frag1 ] the full relations can be found in @xcite . an interesting subleading asymmetry involving @xmath181 is a @xmath232 single spin asymmetry appearing as the structure functions @xmath233 in eq . [ sidiswanti ] for a polarized lepton but unpolarized target @xcite , & & _ lo = _ e y x__b^2 e^a(x__b)h_1^(1)a(z_h ) [ as1 ] where @xmath234 . this cross section involves , besides the time - reversal odd fragmentation function @xmath181 , the distribution function @xmath229 . the tilde function that appear in the cross sections is in fact the socalled interaction dependent part of the twist three functions . it would vanish in any naive parton model calculation in which cross sections are obtained by folding electron - parton cross sections with parton densities . considering the relation for @xmath235 one can state it as @xmath236 = @xmath237 in the absence of quark - quark - gluon correlations . the inclusion of the latter also requires diagrams dressed with gluons as shown in fig . [ figdis1 ] . we have sofar neglected two problems . the first problem is that the correlation function @xmath59 discussed in previous sections involve two quark fields at different space - time points and hence are not color gauge invariant . the second problem comes from the gluonic diagrams similar as the ones we have discussed in the previous section ( see fig . [ figdis1 ] ) we note that diagrams involving matrix elements with longitudinal ( @xmath238 ) gluon fields , @xmath239 do not lead to any suppression . the reason is that because of the @xmath240-index in the gluon field the matrix element is proportional to @xmath241 , @xmath86 or @xmath242 rather than the proportionality to @xmath243 or @xmath164 that we have seen in eq . [ phia ] for a gluonic matrix element with transverse gluons . a straightforward calculation , however , shows that the gluonic diagrams with one or more longitudinal gluons involve matrix elements ( soft parts ) of operators @xmath244 , @xmath245 , @xmath246 , etc . that can be resummed into a correlation function _ ij(x ) = . e^ip p , s_j(0)u(0,)_i ( ) p , s|_^+ = _ = 0 , where @xmath247 is a gauge link operator ( 0 , ) = p(-i_0^^- d^-a^+ ( ) ) ( path - ordered exponential with path along @xmath248-direction ) . et voila , the unsuppressed gluonic diagrams combine into a color gauge invariant correlation function . we note that at the level of operators , one expands ( 0 ) ( ) = _ n ( 0)__1 __n(0 ) , in a set of local operators , but only the expansion of the nonlocal combination with a gauge link ( 0 ) ( ) = _ n ( 0)d__1 d__n(0 ) , is an expansion in terms of local gauge invariant operators . the latter operators are precisely the local ( quark ) operators that appear in the operator product expansion applied to inclusive deep inelastic scattering . for the @xmath163-dependent functions , one finds that inclusion of @xmath238 gluonic diagrams leads to a color gauge invariant matrix element with links running via @xmath249 @xcite . for instance in lepton - hadron scattering one finds ( x , p_t ) = . e^ip p , s(0)u(0 , ) u ( , ) ( ) p , s|_^+ = 0 , where the gauge links are at constant @xmath250 . one can multiply this correlator with @xmath164 and make this into a derivative @xmath251 . including the links one finds the color gauge invariant result & & p_^_ij(x , p _ ) = ( _ ^)_ij(x , p _ ) + & & = e^ip \ { p , s_j(0 ) u(0 , ) id_^_i ( ) p , s|_^+=0 + & & - p , s_j(0)u(0 , ) _ ^^-d^- u ( , ) + & & gg^+ ( ) u(,)_i ( ) p , s|_^+=0 } , which gives after integration over @xmath163 the expected result @xmath252 . note that in @xmath253 gauge all the gauge links disappear , while one has @xmath254 , but there presence is essential to perform the above differentiations . the explicit treatment of transverse momenta provides also a transparent way to include the evolution equations for quark distribution and fragmentation functions . remember that we have assumed that soft parts vanish sufficiently fast as a function of the invariants @xmath82 and @xmath255 , which at constant @xmath256 implies a sufficiently fast vanishing as a function of @xmath257 . this simply turns out not to be true . assuming that the result for @xmath258 is given by the diagram shown in fig . [ figdis2 ] one finds f_1(x , p_^2 ) & = & ( ^2-p_^2)f_1(x , p_^2 ) + & + & ( p_^2-^2 ) _ x^1 p_qq()f_1(y;^2 ) , where @xmath259 and the splitting function is given by p_qq(z ) = c_f , with @xmath260 and the color factor @xmath261 = 4/3 for @xmath262 . with the introduction of the scale in @xmath263 one sees that the scale dependence satisfies = _ x^1 p_qq()f_1(y;^2 ) . this is the standard @xcite nonsinglet evolution equation for the valence quark distribution function . for the flavor singlet combination of quark distributions or the sea distributions one also needs to take into account contributions as shown in fig . [ figdis2 ] ( right ) involving the gluon distribution functions related to matrix elements with gluon fields @xmath264 but otherwise proceeding along analogous lines . the @xmath265-function contribution can be explicitly calculated by including vertex corrections ( socalled virtual diagrams ) , but it is easier to derive them by requiring that the sum rules for @xmath113 remain valid under evolution , which requires that @xmath266 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ except for logarithmic contributions also finite @xmath230 contributions show up in deep inelastic scattering @xcite . for instance in inclusive scattering one finds that the lowest order result for @xmath19 is of this type , @xmath267 , \nonumber \\ & & \end{aligned}\ ] ] the second term involving the gluon distribution function @xmath268 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we will present here the evolution of the transverse moments as they are deduced from known evolution equations for twist-2 functions and for the interaction dependent tilde functions . for the latter we will only exhibit the large @xmath269 results , for which the evolution can be written down in a compact way @xcite as mentioned the evolution of the twist-2 functions and the tilde functions in known . the twist-2 functions have an autonomous evolution of the form f(x , ) = _ x^1 p^[f]()f(y , ) , where @xmath270 = @xmath271 and @xmath272}$ ] are the splitting functions . the leading order results for the non - singlet twist-2 functions ( with the usual + prescription ) @xcite are & & p^[f_1 ] ( ) = p^[g_1 ] ( ) = c_f , + & & p^[h_1 ] ( ) = c_f . in the large @xmath269 limit , also the tilde functions have an autonomous evolution . in leading order one has for the interaction - dependent functions @xcite & & p^[f ] ( ) = , with @xmath273 for @xmath274 , @xmath275 for @xmath276 and @xmath277 for @xmath235 . using these splitting functions and the relations given in the previous section , we find the evolution of the transverse moments , & & g_1t^(1)(x , ) = n_c_x^1 dy\ { g_1t^(1)(y , ) + & & + g_1(y , ) } , + & & h_1l^(1)(x , ) = n_c_x^1 dy\ { h_1l^(1)(y , ) + & & -h_1(y,)}. one can also analyse the fragmentation functions or use some specific reciprocity relations @xcite . furthermore , we note that apart from a @xmath278 matrix the operator structures of the t - odd functions @xmath279 and @xmath280 are in fact the same as those of @xmath217 and @xmath281 ( or as mentioned before , they can be considered as the imaginary part of these functions @xcite ) . with these ingredients one immediately obtains for the non - singlet functions the ( autonomous ) evolution of the t - odd fragmentation functions . in particular we obtain for the collins fragmentation function ( at large @xmath269 ) , & & z h_1^(1)(z , ) = n_c _ z^1 dy y h_1^(1)(y , ) , + & & which should prove useful for the comparison of data on collins function asymmetries from different experiments , performed at different energies . in these lectures i have discussed aspects of hard scattering processes , in particular inclusive and 1-particle inclusive lepton - hadron scattering . the goal is the study of the quark and gluon structure of hadrons . for example , by considering polarized targets or particle production one can measure spin and azimuthal asymmetries and use them to obtain information on specific correlations between spin and momenta of the partons . the reason why this is a promising route is the existence of a field theoretical framework that allows a clean study of the observables as well - defined hadronic matrix elements . roberts , _ the structure of the proton _ , cambridge university press 1990 j.p . ralston and d.e . soper , nucl . phys . * b 152 * ( 1979 ) 109 . tangerman and p.j . mulders , phys . rev . * d 51 * ( 1995 ) 3357 . soper , phys . rev . * d 15 * ( 1977 ) 1141 ; phys . * 43 * ( 1979 ) 1847 . jaffe , nucl . * b 229 * ( 1983 ) 205 . jaffe and x. ji , nucl . * b 375 * ( 1992 ) 527 . x. artru and m. mekhfi , z. phys . * 45 * ( 1990 ) 669 , j.l . cortes , b. pire and j.p . ralston , z. phys . * c 55 * ( 1992 ) 409 . j. soffer , phys . lett . * 74 * ( 1995 ) 1292 . d. boer , phys . rev . * d 62 * ( 2000 ) 094029 ; nucl . * b 603 * ( 2001 ) 195 . n. hammon , o. teryaev and a. schfer , phys . lett . * b 390 * ( 1997 ) 409 ; d. boer , p.j . mulders and o.v . teryaev , phys . rev . * d 57 * ( 1998 ) 3057 ; d. boer and j. qiu , hep - ph/0108179 . d. sivers , phys . rev . * d 41 * ( 1990 ) 83 and phys . rev . * d 43 * ( 1991 ) 261 . m. anselmino , m. boglione and f. murgia , phys . lett . * b 362 * ( 1995 ) 164 ; m. anselmino and f. murgia , phys . lett . * b 442 * ( 1998 ) 470 . a. bacchetta , m. boglione , a. henneman and p.j . mulders , phys . * 85 * ( 2000 ) 712 . a. bacchetta and p.j . mulders , phys . * b 518 * ( 2001 ) 85 . mulders and j. rodrigues , phys . rev . * d 63 * ( 2001 ) 094021 . collins and d.e . soper , nucl . * b 194 * ( 1982 ) 445 . j. collins , nucl . * b 396 * ( 1993 ) 161 . a. de rjula , j.m . kaplan and e. de rafael , nucl . * b 35 * ( 1971 ) 365 ; k. hagiwara , k. hikasa and n. kai , phys . rev . * d 27 * ( 1983 ) 84 . jaffe and x. ji , phys . * 71 * ( 1993 ) 2547 . phys . rev . * d 57 * ( 1998 ) 3057 . j. levelt and p.j . mulders , phys . rev . * d 49 * ( 1994 ) 96 ; phys . * b 338 * ( 1994 ) 357 . d. boer , r. jakob and p.j . mulders , nucl . * b 564 * ( 2000 ) 471 . efremov , o.g . smirnova , l.g . tkachev , in proceedings of the 13 . international symposium on high - energy spin physics ( spin98 ) , protvino , russia , 8 - 12 sep 1998 , nucl . * 74 * ( 1999 ) 49 ; e. efremov , o.g . smirnova and l.g . tkatchev , nucl . phys . proc . * b 79 * ( 1999 ) 554 . a. airapetian _ et al . _ , hermes collaboration , phys . * 84 * ( 2000 ) 4047 . a. bravar , nucl . * b 79 * ( 1999 ) 521 . h. avakian , nucl . * b 79 * ( 1999 ) 523 . a. kotzinian and p.j . mulders , phys . * d 54 * ( 1996 ) 1229 . a. kotzinian , nucl . phys . * b 441 * ( 1995 ) 234 . tangerman and p.j . mulders , phys . * b 352 * ( 1995 ) 129 . bukhvostov , e.a . kuraev and l.n . lipatov , sov . phys . * jetp 60 * ( 1984 ) 22 . mulders and r.d . tangerman , nucl . phys . * b461 * ( 1996 ) 197 and erratum , nucl . phys . * b 484 * ( 1997 ) 538 . d. boer and p.j . mulders , phys . rev . * d 57 * ( 1998 ) 5780 . h. burkhardt and w.n . cottingham , ann . ( n.y . ) * 56 * ( 1976 ) 453 . s. wandzura and f. wilczek , phys . rev . * d 16 * ( 1977 ) 707 . m. boglione and p.j . mulders , phys . rev . * d 60 * ( 1999 ) 054007 . d. boer and p.j . mulders , nucl . * b 569 * ( 2000 ) 505 . a. henneman , d. boer and p.j . mulders , hep - ph/0104271 . see e.g. a.v . belitsky , lectures given at the xxxi pnpi winter school on nuclear and particle physics , st . petersburg , repino , february , 1997 , hep - ph/9703432 . g. altarelli and g. parisi , nucl . * b 126 * ( 1977 ) 298 . f. baldracchini _ _ , fortsch . phys . * 29/30 * ( 1981 ) 505 ; + x. artru and m. mekhfi , z. phys . * c 45 * ( 1990 ) 669 . a. ali , v.m . braun , g. hiller , phys . b 266 * ( 1991 ) 117 ; + i.i . balitsky , v.m . braun , y. koike , k. tanaka , phys . * 77 * ( 1996 ) 3078 ; + x. ji and j. osborne , eur . j. * c 9 * ( 1999 ) 487 .

in these lectures i want to discuss how the structure functions in deep inelastic scattering relate to quark and gluon correlation functions . in particular we will consider the issue of intrinsic transverse momenta of quarks , which becomes important in processes like 1-particle inclusive leptoproduction . some examples of cross sections and asymmetries , in particular in polarized scattering processes are discussed . we also discuss the operator structure for azimuthal asymmetries and their evolution .

j.m.y . acknowledges s. baroni , m. bender , k. hagino and p .- h . heenen for their stimulating discussions and useful comments on this manuscript . this work was partly supported by the nsfc under grant nos . 10947013 , 11105110 and 11105111 , the european union s seventh framework programme ensar under grant agreement n262010 , the fundamental research funds for the central universities ( xdjk2010b007 and xdjk2011b002 ) , and the nsf of chongqing cstc2011jja0376 . d. vretenar , a. v. afanasjev , g. a. lalazissis , and p. ring , phys . rep . * 409 * ( 2005 ) 101 . j. meng , h. toki , s .- zhou , s. q. zhang , w. h. long , and l. s. geng , prog . * 57 * ( 2006 ) 470 .

the possible existence of a bubble " structure in the proton density of @xmath0si has recently attracted a lot of research interest . to examine the existence of the bubble " structure in low - lying states , we establish a relativistic version of configuration mixing of both particle number and angular momentum projected quadrupole deformed mean - field states and apply this state - of - the - art beyond relativistic mean - field method to study the density distribution of the low - lying states in @xmath0si . an excellent agreement with the data of low - spin spectrum and electric multipole transition strengths is achieved without introducing any parameters . we find that the central depression in the proton density is quenched by dynamic quadrupole shape fluctuation , but not as significantly as what has been found in a beyond non - relativistic mean - field study . our results suggest that the existence of proton bubble " structure in the low - lying excited @xmath1 and @xmath2 states is very unlikely . covariant density functional theory , beyond mean - field approximation , low - lying states , density distribution _ introduction_.@xmath3 in recent years , there has been a growing interest in searching for bubble " nuclei , in which the density in the center vanishes or is significantly lower than saturation density . the bubble " nuclei are of particular interest because they have different mean - field potentials from those of normal nuclei with fermi - type density distribution . in particular , the splitting of spin - orbit partners located mainly at the nuclear center is reduced in the bubble " nuclei . some modern parametrizations of the relativistic mean field ( rmf ) @xcite and of the skyrme energy density functional ( edf ) @xcite , as well as the mean - field calculation using semi - realistic interactions @xcite predict a hollow proton density for @xmath4 and some neutron - rich ar isotopes . at the time being , @xmath4 stands out as the only candidate on which many different studies agree . the possible proton `` bubble '' structure of this nucleus has also been suggested as an explanation for the recently observed reduction of the spin - orbit splitting using the transfer reactions @xmath5 and @xmath6 @xcite . recently , the intruder @xmath1 state and its weak electric monopole transition to the ground - state were measured @xcite . the results support the coexistence structure of a spherical ground - state with a large deformed @xmath1 state in @xmath0si . the spectroscopy of low - lying states provides a strong test of the nuclear structure models that have been used to study the density profiles . most recently , the stability of bubble " structure against dynamical effects in @xmath0si has been examined in the framework of a particle - number ( pn ) and angular - momentum ( am ) projected generator coordinate method ( gcm ) based on hartree - fock - bogoliubov ( hfb ) states with axial quadrupole deformation using the non - relativistic skyrme force sly4 @xcite . it has been shown that the dynamic effect of quadrupole shape fluctuation significantly altered the radial density profile , and brought it closer to a fermi - type density distribution . we noted , however , that the spectroscopic properties of the observed low - lying states were not reproduced very well @xcite . the calculated two lowest @xmath7 states were too strongly mixed , which might overestimate the effect of shape mixing on the density of ground - state . moreover , in the context of projected gcm based on self - consistent mean - field approaches , the density profiles of excited states have not been studied yet . the existence of the proton bubble " structure in the low - lying excited states of @xmath0si is not known . during the past decades , the rmf theory has achieved great success in describing many nuclear phenomena for both stable and exotic nuclei over the entire nuclear chart with a few universal parameters @xcite . it incorporates many important relativistic effects , such as the presence of large lorentz scalar and vector fields with approximately equal magnitude and opposite sign . this leads to a new saturation mechanism via the difference between the scalar and vector densities , and naturally to the large spin - orbit interaction needed for the understanding of magic numbers in finite nuclei . the aim of the present letter is to address the above issue using a beyond rmf approach . _ the method_.@xmath3 to this end , we extend the beyond rmf method presented in ref . @xcite by including additionally pn projection and restricting it to axially deformed states by imposing triaxial deformation parameter @xmath8 to be @xmath9 or @xmath10 . in this case , the wave function of nuclear low - lying state is given by the superposition of both pn and am projected rmf wave functions constrained to have different intrinsic axial deformations @xmath11 , @xmath12 with @xmath13 , @xmath14 , @xmath15 being the projection operators onto good number of angular momentum , neutrons and protons , respectively . the weight factors @xmath16 and the energies @xmath17 of the states @xmath18 are the solutions of the hill - wheeler - griffin equation @xcite @xmath19 \ , f_\alpha^{jnz}(\beta ' ) = 0,\ ] ] where @xmath20 and @xmath21 are the norm kernel and the energy kernel , respectively . in the calculations , the energy overlap in the energy kernel is taken with the same functional form as the nuclear mean - field energy , but replacing the densities and currents with mixed ones @xcite . we note that in recent years several similar beyond mean - field methods ( gcm+pnp+amp ) are developed by different groups based on either non - relativistic or relativistic edfs @xcite . once the wavefunctions of low - lying states are found , it is straightforward to calculate the density distribution in @xmath22-space . the density distribution for the state @xmath18 is found as @xcite @xmath23 where the @xmath24 is defined as @xmath25 the sum over index @xmath26 represents summation over all occupied single - particle states for neutrons or protons . the reduced electric quadrupole ( @xmath27 ) transition strengths between low - lying states are given by @xmath28 and are calculated directly in the laboratory frame without approximation . the reduced matrix elements in the above expression are determined as @xmath29 where @xmath30 is the electric quadrupole moment operator . the nuclear matrix element entering the electric monopole decay from @xmath31 to @xmath32 through the emission of conversion electrons is determined by @xmath33 where @xmath34 is an index running over all proton single - particle states . the radius @xmath35 is given by @xmath36 fm . since the electric transition matrix elements are calculated in the full model space of occupied single - particle states , there is no need to introduce effective charges , and the bare charge for protons is used instead . _ numerical details_.@xmath3 in the rmf calculations , parity , @xmath37-simplex symmetry , and time - reversal invariance are imposed . the dirac equation is solved by expanding in the basis of eigenfunctions of a three - dimensional harmonic oscillator in cartesian coordinate with 10 major shells , which are found to be sufficient to obtain reasonably converged results for @xmath0si . a constraint on the axial mass quadrupole moment @xmath38 is used to generate mean - field states @xmath39 with different intrinsic deformations @xmath40 . the point - coupling type of relativistic effective force pc - pk1 @xcite is adopted . pairing correlations between nucleons are treated with the bcs approximation using a density - independent @xmath41 force with a smooth cutoff factor @xcite . the strength parameters of the pairing force are @xmath42 and @xmath43 mev@xmath44@xmath45 for neutrons and protons respectively . the gauss - legendre quadrature is used for integrals over the euler angle @xmath46 and gauge angle @xmath47 in the calculations of the am+pn projected norm and hamiltonian kernels . the number of mesh points in the interval @xmath48 $ ] for @xmath46 and @xmath49 are chosen as 14 and 9 respectively . the pfaffian method @xcite is implemented to calculate the norm overlap , the phase of which can be uniquely determined in this way . si . the electric monopole ( @xmath50 ) transition between the first two @xmath7 states is indicated with the value of @xmath51 . the electric quadrupole ( @xmath27 ) transition is indicated with the value in units of @xmath52 @xmath53 . , width=321 ] _ results and discussions_.@xmath3 figure [ spectrum ] displays the low - lying states in @xmath0si from the beyond mean - field calculations using both the non - relativistic sly4 force and the relativistic pc - pk1 force , in comparison with the recently measured data @xcite . as we have discussed in ref . @xcite that the calculations using the sly4 force reproduced the energy of the recently observed low - energy @xmath1 state and the interband @xmath54 value rather well . however , the electric monopole transition strength @xmath55 and the in - band @xmath56 value were overestimated , which indicates that the two lowest @xmath7 gcm states are too strongly mixed in the calculation . in the relativistic calculations using the pc - pk1 force , however , both the excitation energies and @xmath50 , @xmath27 transition strengths of the low - lying states are reproduced quite well , which gives us a strong motivation to study the density distributions of low - lying states using the wave functions obtained in the relativistic calculations . ) ] of the @xmath57 states in @xmath0si.,width=302 ] figure [ gcmwfs ] displays the distribution of the collective wave functions @xmath58 in deformation @xmath11 degree of freedom for the @xmath57 states , among which the electric multipole transitions have been observed @xcite , where the @xmath59 are related to the weight function @xmath60 by the following relation , @xmath61 and are orthonormal to each other . the shapes of collective wave functions for the @xmath62 and @xmath1 states are similar to those obtained in ref . @xcite from the calculations using the sly4 force . however , there are differences in detail . the wave function of ground - state is more concentrated around the spherical shape in the relativistic case , while that of @xmath1 state has more weight in large deformed configurations ( dominated by the intruder neutron @xmath63 orbital ) . as a consequence , compared with the sly4 results , the @xmath64 and @xmath65 values are reduced and become closer to the data in the beyond rmf calculations . the density distribution of protons in @xmath0si from the rmf calculation is displayed in fig . [ density ] , which shows an evident central depletion . this density distribution corresponds to the spherical mean - field configuration , for which case , the occupation probability of proton @xmath66 orbital is zero due to pairing collapse . the mixing of configurations of different intrinsic shapes resulting from dynamic fluctuation in quadrupole deformations can alter the occupancy of the @xmath66 orbital and therefore change the density distribution in the center . figure [ compdens ] displays the comparison of mean - field and beyond mean - field ( gcm+pnp+amp ) calculated density distributions for the ground - state in @xmath0si . it shows that the quenching of the depletion of proton density in the center by shape mixing is a common feature in the beyond mean - field calculations . however , the beyond rmf calculation gives a smaller quenching effect than that the one found in the beyond non - relativistic mean - field calculation , which can be mainly attributed to the difference in the distribution of wave function in the collective space , i.e. , the wave function of ground - state in this work is more concentrated around the spherical shape than that given in ref . @xcite . -@xmath67 plane at @xmath68 fm from the rmf calculation using the pc - pk1 force for @xmath0si ( in @xmath69).,width=302 ] si . the non - relativistic results , taken from ref . @xcite , are given for comparison.,width=264 ] si and @xmath70s from the relativistic calculations using the pc - pk1 force . the experimental data for @xmath70s are taken from ref . @xcite , width=264 ] . [ charge - density ] figure [ charge - density ] displays the comparison of charge density distributions in @xmath0si and @xmath70s from the relativistic calculations using the pc - pk1 force . the charge density is calculated by convolution of the proton density with a gaussian form factor with a proton size @xmath71 fm , which for spherically symmetric density distributions leads to @xcite @xmath72 \ , .\ ] ] the charge density of mean - field result ( @xmath73 ) for @xmath70s is much higher than the experimental data . this phenomenon is also observed in other rmf calculations using the nl3 and dd - me2 forces @xcite . figure [ charge - density ] demonstrates that after including the dynamic correlation effects in the projected gcm calculation , the charge density of @xmath70s is in excellent agreement with the data , which gives us confidence in the prediction for the charge density of @xmath0si . in the mean - field approaches , a central depletion of the proton density in @xmath0si could induce a non - zero density derivative in the center and thus reduce the strength of the spin - orbit interaction for the inner orbits . this has been suggested as an explanation for the reduction of neutron @xmath74 splitting between @xmath75s and @xmath76si inferred from transfer reactions @xcite . to examine this effect at the mean - field level , we plot in fig . [ spe ] the single - particle energy spectra for protons and neutrons corresponding to the spherical configuration of @xmath0si and @xmath70s from the rmf calculations using the pc - pk1 force . as expected , the splitting of spin - orbit doublets @xmath74 is reduced significantly from 2.77(2.66 ) mev to 0.73(0.80 ) mev for protons ( neutrons ) when going from @xmath70s to @xmath0si . it should be noted that the spin - orbit interaction is emerging naturally from the derivative of vector and scalar fields in the rmf approaches and no adjustable parameter is introduced . however , one should not compare these values directly with those inferred from transfer reactions , as discussed in ref . @xcite . moreover , as shown in fig . [ charge - density ] that the pure rmf calculation overestimates the central density in @xmath70s , and therefore enhances the spin - orbital splitting . furthermore , the neutron @xmath77 level in @xmath0si is weakly bound at -0.063 mev , whereas the neutron @xmath78 level in @xmath0si and @xmath70s is unbound at + 0.74 mev and + 0.48 mev respectively , in which case , the coupling to the continuum has to be carefully taken into account , which is beyond the scope of present study . figure [ gcmdens ] displays the density distributions of neutrons , protons and charges for the first two @xmath7 states , in comparison with the mean - field calculated results of pure spherical configuration . it shows clearly that the densities of both neutrons and protons in the ground - state are closer to the fermi - type distribution in the projected gcm calculation , in comparison with the pure rmf calculation ( @xmath73 ) . in particular , for the @xmath1 state , the central depression in the density distribution is not visible . si and @xmath70s from the rmf calculations using the pc - pk1 force . the size of spin - orbit splitting is indicated with the value in units of mev . , width=302 ] states ( solid lines ) , in comparison with the mean - field calculated results for the spherical configuration ( dashed lines).,width=264 ] ) of the @xmath2 state in @xmath0si . the @xmath79 component is scaled by dividing a factor of @xmath80.,width=302 ] for non - zero spin states , the density is deformed , in which case , it is convenient to define a radial ( reduced ) transition density with angular momentum transfer @xmath81 @xcite , @xmath82 where @xmath83 is a clebsch - gordan coefficient . we note that the integration of the radial transition density @xmath84 , multiplied with @xmath85 over the radial coordinate @xmath22 is related to the mean - square radius of the state @xmath86 by the relation @xmath87 . we find that the mean - square radius of the first three low - lying states are respectively @xmath88 fm , @xmath89 fm , and @xmath90 fm . moreover , the spectroscopic quadrupole moment of @xmath2 state is 9.48 @xmath91 @xmath92 , which indicates that the @xmath2 state is predominately oblate deformed . figure [ densj2 ] displays the radial transition densities ( @xmath93 ) of the @xmath2 state in @xmath0si . it is clearly shown that the central depression does not exist in the @xmath79 component of radial transition density . the disappearance of bubble " structure in the @xmath1 and @xmath2 states is due to the mixing of ( prolate ) deformed configurations , in which , the component originated from the proton spherical @xmath66 orbital has a sizable occupation probability . l|cccc|cc state & @xmath94 & @xmath95 & @xmath96 & @xmath97 & @xmath98 & @xmath99 + @xmath73 & 0.056 & 0.077 & 0.27 & 0.15 & 0.14 & 0.04 + @xmath62 & 0.059 & 0.076 & 0.22 & 0.11 & 0.11 & 0.01 + @xmath1 & 0.071 & 0.074 & 0.04 & @xmath30.08 & 0.01 & @xmath30.10 + @xmath2 ( @xmath79 ) & 0.073 & 0.075 & 0.03 & @xmath30.11 & & + table [ factor ] lists the values of the central and maximal proton densities , and the depletion factors @xmath100 and @xmath101 @xcite for the proton and charge ( ch . ) densities in @xmath0si from the calculations using the pc - pk1 force , where @xmath102 with @xmath103 being the saturation value of the proton density , @xmath104 @xmath105 @xmath69 for @xmath4 . the beyond mean - field effects , dominated by the shape mixing , reduce the depletion factor @xmath96 from 0.14 to 0.11 and @xmath97 from 0.04 to 0.01 for the charge density , which are compared with the results of the sly4 force , i.e. , @xmath96 from 0.16 to 0.09 and @xmath97 from 0.09 to 0.04 . for the observed low - lying @xmath1 and @xmath2 states , the depletion factor is close to zero , which demonstrates the disappearance of proton bubble " structure quantitatively . finally , we point out that the implementation of lipkin - nogami ( ln ) prescription for pairing correlation in our rmf calculations would bump protons onto @xmath66 orbital and consequently reduce the bubble " structure further . _ summary and conclusions_.@xmath3 we have established a beyond rmf method by mixing of both particle number and angular momentum projected quadrupole deformed states and applied it to study the excitation energies , electric multipole transition strengths and the density distributions for the low - lying states in @xmath0si . the recently observed spectroscopic data have been reproduced quite well without introducing any parameters , which provides us confidence on the reliability of the wave functions and the resultant density distributions . moreover , we have found that for @xmath70s , the dynamic correlation effects have significant influence on the central charge density . after considering these effects in the beyond rmf calculation , the charge density distribution is in excellent agreement with the data , which gives us more confidence in the prediction for the charge density of @xmath0si . we have found that the central depression in the proton and charge densities of @xmath0si is quenched by dynamic quadrupole shape fluctuation , but not as significantly as what has been found in a non - relativistic study using the sly4 force . our studies suggest that the existence of a proton bubble " structure in the low - lying excited @xmath1 and @xmath2 states is very unlikely . these findings are hoped to be examined in the new generation of electron - rib colliders scrit ( self confining radioactive isotope target in japan @xcite ) and elise ( electron - ion scattering in a storage ring in germany @xcite ) in the near future .

a number of papers have recently considered the problem of constructing a coherent quantum observer for a quantum system ; e.g. , see @xcite . in the coherent quantum observer problem , a quantum plant is coupled to a quantum observer which is also a quantum system . the quantum observer is constructed to be a physically realizable quantum system so that the system variables of the quantum observer converge in some suitable sense to the system variables of the quantum plant . the papers @xcite considered the problem of constructing a direct coupling quantum observer for a given closed quantum system . in @xcite , the proposed observer is shown to be able to estimate some but not all of the plant variables in a time averaged sense . also , the paper @xcite shows that a possible experimental implementation of the augmented quantum plant and quantum observer system considered in @xcite may be constructed using a non - degenerate parametric amplifier ( ndpa ) which is coupled to a beamsplitter by suitable choice of the ndpa and beamsplitter parameters . one important limitation of the direct coupled quantum observer results given in @xcite is that both the quantum plant and the quantum observer are closed quantum systems . this means that it not possible to make an experimental measurement to verify the properties of the quantum observer . in this paper , we address this difficulty by extending the results of @xcite to allow for the case in which the quantum observer is an open quantum linear system whose output can be monitored using homodyne detection . in this case , it is shown that similar results can be obtained as in @xcite except that now the observer output is subject to a white noise perturbation . however , by suitably designing the observer , it is shown that the level of this noise perturbation can be made arbitrarily small ( at the expense of slow observer convergence ) . also , the results of @xcite are extended to show that a possible experimental implementation of the augmented quantum plant and quantum observer system may be constructed using a non - degenerate parametric amplifier ( ndpa ) which is coupled to a beamsplitter by suitable choice of the ndpa and beamsplitter parameters . in this case , the ndpa contains an extra field channel as compared to the result in @xcite and this extra channel is used for homodyne detection in the observer . in this section , we extend the theory of @xcite to the case of a direct coupled quantum observer which is also coupled to a field to enable measurements to be made on the observer . in our proposed direct coupled coherent quantum observer , the quantum plant is a single quantum harmonic oscillator which is a linear quantum system ( e.g. , see @xcite ) described by the non - commutative differential equation @xmath0 where @xmath1 denotes the system variable to be estimated by the observer and @xmath2 . this quantum plant corresponds to a plant hamiltonian @xmath3 . here @xmath4 $ ] where @xmath5 is the plant position operator and @xmath6 is the plant momentum operator . it follows from ( [ plant ] ) that the plant system variables @xmath7 will remain fixed if the plant is not coupled to the observer . we now describe the linear quantum system which will correspond to the quantum observer ; see also @xcite . this system is described by a quantum stochastic differential equation ( qsde ) of the form @xmath8 where @xmath9 $ ] is a @xmath10 vector of quantum noises expressed in quadrature form corresponding to the input field for the observer and @xmath11 is the corresponding output field ; e.g. , see @xcite . the observer output @xmath12 will be a real scalar quantity obtained by applying homodyne detection to the observer output field . @xmath13 , @xmath14 , @xmath15 . also , @xmath16 $ ] is a vector of self - adjoint system variables corresponding to the observer position and momentum operators ; e.g. , see @xcite . we assume that the plant variables commute with the observer variables . the system dynamics ( [ observer ] ) are determined by the observer system hamiltonian and coupling operators which are operators on the underlying hilbert space for the observer . for the quantum observer under consideration , this hamiltonian is a self - adjoint operator given by the quadratic form : @xmath17 , where @xmath18 is a real symmetric matrix . also , the coupling operator @xmath19 is defined by a matrix @xmath20 so that @xmath21 = w_o x_o.\ ] ] then , the corresponding matrices @xmath22 , @xmath23 and @xmath24 in ( [ observer ] ) are given by @xmath25 where @xmath26;\ ] ] e.g. , see @xcite . furthermore , we will assume that the quantum observer is coupled to the quantum plant as shown in figure [ f1 ] . in addition , we define a coupling hamiltonian which defines the coupling between the quantum plant and the quantum observer : @xmath27 the augmented quantum linear system consisting of the quantum plant and the quantum observer is then a linear quantum system described by the total hamiltonian @xmath28 where @xmath29,~~ r_a = \left[\begin{array}{ll}0 & r_{c } \\ r_{c}^t & r_{o } \end{array}\right],\ ] ] and the coupling operator @xmath19 defined in ( [ wo ] ) . extending the approach used in @xcite , we assume that we can write @xmath30 @xmath31 , @xmath32 where @xmath33 , @xmath34 , @xmath35 and @xmath36 . in addition , we assume @xmath37 then , we can write the qsdes describing the closed loop system as follows : @xmath38x_o dt + 2j\beta \alpha^t x_p dt - \sqrt{\kappa}dw ; \nonumber \\ dy_o & = & w_ox_o dt + dw \nonumber \\ & = & \sqrt{\kappa } x_o dt + dw;\end{aligned}\ ] ] e.g. , see @xcite . now it follow from ( [ plant ] ) and ( [ alpha ] ) that @xmath39 hence , it follows from the first equation in ( [ augmented1 ] ) that @xmath40 that is , the quantity @xmath41 remains constant even after the quantum plant is coupled to the quantum observer . in addition , we can re - write the remaining equations in ( [ augmented1 ] ) as @xmath42x_o dt + 2j\beta z_p dt - \sqrt{\kappa}dw ; \nonumber \\ dy_o & = & \sqrt{\kappa } x_o dt + dw;\end{aligned}\ ] ] to analyse the system ( [ augmented2 ] ) , we first calculate the steady state value of the quantum expectation of the observer variables as follows : @xmath43^{-1 } j\beta z_p \nonumber \\ & = & \frac{4}{\kappa^2 + 16 \omega_o^2}\left[\begin{array}{ll } \kappa & 4 \omega_o \\-4\omega_o & \kappa \end{array}\right]j\beta z_p.\end{aligned}\ ] ] then , we define the quantity @xmath44j\beta z_p.\ ] ] we can now re - write the equations ( [ augmented2 ] ) in terms of @xmath45 as follows @xmath46x_o dt + 2j\beta z_p dt - \sqrt{\kappa}dw \nonumber \\ & = & \left[\begin{array}{ll } -\frac{\kappa}{2 } & 2 \omega_o \\-2\omega_o & -\frac{\kappa}{2 } \end{array}\right]\tilde x_o dt \nonumber \\ & & - 2 \left[\begin{array}{ll } -\frac{\kappa}{2 } & 2 \omega_o \\-2\omega_o & -\frac{\kappa}{2 } \end{array}\right ] \left[\begin{array}{ll } -\frac{\kappa}{2 } & 2 \omega_o \\-2\omega_o & -\frac{\kappa}{2 } \end{array}\right]^{-1 } j\beta z_p dt\nonumber \\ & & + 2j\beta z_p dt - \sqrt{\kappa}dw \nonumber \\ & = & \left[\begin{array}{ll } -\frac{\kappa}{2 } & 2 \omega_o \\-2\omega_o & -\frac{\kappa}{2 } \end{array}\right]\tilde x_o dt - \sqrt{\kappa}dw ; \nonumber \\ dy_o & = & \sqrt{\kappa } \tilde x_o dt - 2 \sqrt{\kappa}\left[\begin{array}{ll } -\frac{\kappa}{2 } & 2 \omega_o \\-2\omega_o & -\frac{\kappa}{2 } \end{array}\right]^{-1 } j\beta z_p dt+ dw \nonumber \\ & = & - 2 \sqrt{\kappa}\left[\begin{array}{ll } -\frac{\kappa}{2 } & 2 \omega_o \\-2\omega_o & -\frac{\kappa}{2 } \end{array}\right]^{-1 } j\beta z_p dt + dw^{out}\end{aligned}\ ] ] where @xmath47 we now look at the transfer function of the system @xmath48\tilde x_o dt - \sqrt{\kappa}dw ; \nonumber \\ dw^{out } & = & \sqrt{\kappa } \tilde x_o dt + dw\end{aligned}\ ] ] which is given by @xmath49^{-1}.\ ] ] it is straightforward to verify that this transfer function is such that @xmath50 for all @xmath51 . that is @xmath52 is all pass . also , the matrix @xmath53 $ ] is hurwitz and hence , the system ( [ augmented4 ] ) will converge to a steady state in which @xmath54 represents a standard quantum white noise with zero mean and unit intensity . hence , at steady state , the equation @xmath55^{-1 } j\beta z_p dt + dw^{out}\ ] ] shows that the output field converges to a constant value plus zero mean white quantum noise with unit intensity . we now consider the construction of the vector @xmath56 defining the observer output @xmath57 . this vector determines the quadrature of the output field which is measured by the homodyne detector . we first re - write equation ( [ yo ] ) as @xmath58 where @xmath59^{-1 } j\beta\ ] ] is a vector in @xmath60 . then @xmath61 hence , we choose @xmath56 such that @xmath62 and therefore @xmath63 where @xmath64 will be a white noise process at steady state with intensity @xmath65 . thus , to maximize the signal to noise ratio for our measurement , we wish to choose @xmath56 to minimize @xmath65 subject to the constraint ( [ kconstraint ] ) . note that it follows from ( [ kconstraint ] ) and the cauchy - schwartz inequality that @xmath66 and hence @xmath67 however , if we choose @xmath68 then ( [ kconstraint ] ) is satisfied and @xmath69 . hence , this value of @xmath56 must be the optimal @xmath56 . we now consider the special case of @xmath70 . in this case , we obtain @xmath71 j\beta = \frac{4}{\sqrt{\kappa}}j\beta.\ ] ] hence , as @xmath72 , @xmath73 and therefore @xmath74 . this means that we can make the noise level on our measurement arbitrarily small by choosing @xmath36 sufficiently small . however , as @xmath75 gets smaller , the system ( [ augmented4 ] ) gets closer to instability and hence , takes longer to converge to steady state . in this section , we describe one possible experimental implementation of the plant - observer system given in the previous section . the plant - observer system is a linear quantum system with hamiltonian @xmath76 and coupling operator defined so that @xmath77 = w_o x_o.\ ] ] furthermore , we assume that @xmath78 , @xmath31 , @xmath32 where @xmath33 , @xmath34 , @xmath35 and @xmath36 . in order to construct a linear quantum system with a hamiltonian of this form , we consider an ndpa coupled to a beamsplitter as shown schematically in figure [ f2 ] ; e.g. , see @xcite . a linearized approximation for the ndpa is defined by a quadratic hamiltonian of the form @xmath79 where @xmath80 is the annihilation operator corresponding to the first mode of the ndpa and @xmath81 is the annihilation operator corresponding to the second mode of the ndpa . these modes will be assumed to be of the same frequency but with a different polarization with @xmath80 corresponding to the quantum plant and @xmath81 corresponding to the quantum observer . also , @xmath82 is a complex parameter defining the level of squeezing in the ndpa and @xmath83 corresponds to the detuning frequency of the @xmath81 mode in the ndpa . the @xmath80 mode in the ndpa is assumed to be tuned . in addition , the ndpa corresponds to a vector of coupling operators @xmath84 $ ] . here @xmath85 , @xmath86 , @xmath87 are scalar parameters determined by the reflectance of the mirrors in the ndpa . from the above hamiltonian and coupling operators , we can calculate the following quantum stochastic differential equations ( qsdes ) describing the ndpa : @xmath88 & = & \left[\begin{array}{ll } 0 & \frac{\epsilon}{2 } \\ \frac{\epsilon}{2 } & 0 \end{array } \right]\left[\begin{array}{l } a^ * \\ b^ * \end{array } \right]dt \nonumber \\ & & - \left[\begin{array}{ll } \frac{\gamma_1}{2 } & 0 \\ 0 & \frac{\gamma_2}{2}+\half\imath \omega_o \end{array } \right]\left[\begin{array}{l } a \\ b \end{array } \right]dt \nonumber \\ & & - \left[\begin{array}{lll } \sqrt{\kappa_1 } & 0 & 0\\ 0 & \sqrt{\kappa_2 } & \sqrt{\kappa_3 } \end{array } \right]\left[\begin{array}{l } da \\ db_1 \\ db_2 \end{array } \right];\nonumber \\ \left[\begin{array}{l } da^{out } \\ db_1^{out}\\ db_2^{out } \end{array } \right ] & = & \left[\begin{array}{ll } \sqrt{\kappa_1 } & 0 \\ 0 & \sqrt{\kappa_2 } \\ 0 & \sqrt{\kappa_3}\end{array } \right ] \left[\begin{array}{l } a \\ b \end{array } \right]dt + \left[\begin{array}{l } da \\ db_1\\ db_2 \end{array } \right];\nonumber \\\end{aligned}\ ] ] where @xmath89 and @xmath90 . we now consider the equations defining the beamsplitter @xmath91 = \left[\begin{array}{ll } \cos \theta & e^{-\imath \phi } \sin \theta \\ -e^{\imath \phi } \sin \theta & \cos \theta \end{array } \right]\left[\begin{array}{l } a^{out } \\ b_1^{out } \end{array } \right]\ ] ] where @xmath92 and @xmath93 are angle parameters defining the beamsplitter ; e.g. , see @xcite . this implies @xmath94 = \left[\begin{array}{ll } \cos \theta & -e^{-\imath \phi } \sin \theta \\ e^{\imath \phi } \sin \theta & \cos \theta \end{array } \right ] \left[\begin{array}{l } a \\ b_1 \end{array } \right].\ ] ] substituting this into the second equation in ( [ qsde ] ) , we obtain @xmath95 \left[\begin{array}{l } da \\ db_1 \end{array } \right ] } \nonumber \\ & & = \left[\begin{array}{ll } \sqrt{\kappa_1 } & 0\\ 0 & \sqrt{\kappa_2 } \end{array } \right ] \left[\begin{array}{l } a \\ b \end{array } \right]dt + \left[\begin{array}{l } da \\ db_1 \end{array } \right ] \end{aligned}\ ] ] and hence @xmath96 \left[\begin{array}{l } da \\ db_1 \end{array } \right]\\ = \left[\begin{array}{ll } \sqrt{\kappa_1 } & 0\\ 0 & \sqrt{\kappa_2 } \end{array } \right ] \left[\begin{array}{l } a \\ b \end{array } \right]dt.\end{aligned}\ ] ] we now assume that @xmath97 . it follows that we can write @xmath98 = } \nonumber \\ & & \frac{1}{2(1-\cos \theta)}\left[\begin{array}{ll } \cos \theta -1 & e^{-\imath \phi } \sin \theta \\ -e^{\imath \phi } \sin \theta & \cos \theta -1 \end{array } \right]\nonumber \\ & & \times \left[\begin{array}{ll } \sqrt{\kappa_1 } & 0\\ 0 & \sqrt{\kappa_2 } \end{array } \right]\left[\begin{array}{l } a \\ b \end{array } \right]dt.\end{aligned}\ ] ] substituting this into the first equation in ( [ qsde ] ) , we obtain @xmath99=\nonumber \\ & & \left[\begin{array}{ll } 0 & \frac{\epsilon}{2 } \\ \frac{\epsilon}{2 } & 0 \end{array } \right]\left[\begin{array}{l } a^ * \\ b^ * \end{array } \right]dt \nonumber \\ & & - \left[\begin{array}{ll } \frac{\gamma_1}{2 } & 0 \\ 0 & \frac{\gamma_2}{2}+\half\imath \omega_o \end{array } \right]\left[\begin{array}{l } a \\ b \end{array } \right]dt \nonumber \\ & & - \frac{1}{2(1-\cos \theta)}\left[\begin{array}{ll } \sqrt{\kappa_1 } & 0\\ 0 & \sqrt{\kappa_2 } \end{array } \right ] \left[\begin{array}{ll } \cos \theta -1 & e^{-\imath \phi } \sin \theta \\ -e^{\imath \phi } \sin \theta & \cos \theta -1 \end{array } \right]\nonumber \\ & & \times \left[\begin{array}{ll } \sqrt{\kappa_1 } & 0\\ 0 & \sqrt{\kappa_2 } \end{array } \right ] \left[\begin{array}{l } a \\ b \end{array } \right]dt \nonumber \\ & & -\left[\begin{array}{l } 0\\ \sqrt{\kappa_3 } \end{array } \right]db_2 ; \nonumber \\ & & db_2^{out } = \sqrt{\kappa_3}b + db_2.\end{aligned}\ ] ] these qsdes can be written in the form @xmath100 & = & f \left[\begin{array}{l } a \\ b \\a^ * \\ b^*\end{array } \right]dt + g \left[\begin{array}{l}db_2\\db_2^*\end{array } \right ] ; \nonumber \\ \left[\begin{array}{l}db_2^{out}\\db_2^{out*}\end{array } \right ] & = & h\left[\begin{array}{l } a \\ b \\a^ * \\ b^*\end{array } \right]dt + \left[\begin{array}{l}db_2\\db_2^*\end{array } \right]\end{aligned}\ ] ] where the matrix @xmath101 is given by @xmath102 $ ] and @xmath103,\\ f_2 & = & \left[\begin{array}{ll}0 & \frac{\epsilon}{2}\\\frac{\epsilon}{2 } & 0\end{array}\right].\end{aligned}\ ] ] also , the matrix @xmath104 is given by @xmath105,\ ] ] and the matrix @xmath106 is given by @xmath107.\ ] ] it now follows from the proof of theorem 1 in @xcite that we can construct a hamiltonian for this system of the form @xmath108m\left[\begin{array}{l } a \\ b \\a^ * \\ b^*\end{array } \right]\ ] ] where the matrix @xmath109 is given by @xmath110 and @xmath111 $ ] . then , we calculate @xmath112 $ ] where @xmath113,\nonumber \\ m_2 & = & \frac{\imath}{2}\left[\begin{array}{ll}0 & \epsilon\\\epsilon & 0\end{array}\right].\end{aligned}\ ] ] also , we can construct the coupling operator for this system in the form @xmath114\left[\begin{array}{l } a \\ b \\a^ * \\ b^*\end{array } \right]\ ] ] where the matrix @xmath115 $ ] is given by @xmath116 hence , @xmath117 , n_2 = 0.\ ] ] we now wish to calculate the hamiltonian @xmath118 in terms of the quadrature variables defined such that @xmath119 = \phi \left[\begin{array}{l } q_p \\ p_p \\q_o \\ p_o\end{array } \right]\ ] ] where the matrix @xmath120 is given by @xmath121.\ ] ] then we calculate @xmath122r\left[\begin{array}{l } q_p \\ p_p \\q_o \\ p_o\end{array } \right]\nonumber \\ & = & \half \left[\begin{array}{ll}x_p\trp & x_o\trp \end{array } \right ] r\left[\begin{array}{l}x_p \\ x_o \end{array } \right]\end{aligned}\ ] ] where the matrix @xmath123 is given by @xmath124,\end{aligned}\ ] ] @xmath125\ ] ] and @xmath126 . hence , @xmath127 comparing this with equation ( [ rc ] ) , we require that @xmath128 = \alpha\beta^t\ ] ] and the condition ( [ kconstraint ] ) to be satisfied in order for the system shown in figure [ f2 ] to provide an implementation of the augmented plant - observer system . we first observe that the matrix on the right hand side of equation ( [ rc1 ] ) is a rank one matrix and hence , we require that @xmath129 = |\delta|^2 - |\epsilon|^2 = 0.\ ] ] that is , we require that @xmath130 note that the function @xmath131 takes on all values in @xmath132 for @xmath133 and hence , this condition can always be satisfied for a suitable choice of @xmath92 . this can be seen in figure [ f4 ] which shows a plot of the function @xmath134 . .,width=302 ] furthermore , we will assume without loss of generality that @xmath135 and hence we obtain our first design equation @xmath136 in practice , this ratio would be chosen in the range of @xmath137 in order to ensure that the linearized model which is being used is valid . we now construct the vectors @xmath138 and @xmath139 so that condition ( [ rc1 ] ) is satisfied . indeed , we let @xmath140,~~ \beta= \left[\begin{array}{l } -\im(\epsilon ) -\im(\delta)\\\re(\epsilon ) + \re(\delta)\end{array}\right].\ ] ] for these values of @xmath138 and @xmath139 , it is straightforward to verify that ( [ rc1 ] ) is satisfied provided that @xmath141 . with this value of @xmath139 , we now calculate the quantity @xmath142 defined in ( [ e ] ) as follows : @xmath143j\beta \\ & = & -\frac{4}{\kappa_3 ^ 2 + 16 \omega_o^2}\left[\begin{array}{ll } \kappa_3 & 4 \omega_o \\-4\omega_o & \kappa_3 \end{array}\right ] \left[\begin{array}{l}\re(\epsilon ) + \re(\delta ) \\\im(\epsilon)+ \im(\delta)\end{array}\right ] . \ ] ] then , the vector @xmath56 defining the quadrature measured by the homodyne detector is constructed according to the equation ( [ k ] ) . in the special case that @xmath144 , this reduces to @xmath145.\ ] ] in terms of complex numbers @xmath146 , we can write this as @xmath147 then , in terms of complex numbers @xmath148 , the formula ( [ k ] ) becomes @xmath149 where @xmath150 denotes complex conjugate . also , as noted in section [ sec : observer ] , the steady state measurement noise intensity is given by @xmath151 which approaches zero as @xmath152 . however , this is at the expense of increasingly slower convergence to steady state . in this paper , we have shown that a direct coupling observer for a linear quantum system can be implemented in the case that the observer can be measured using a homodyne detection measurement . this would allow the plant observer system to be constructed experimentally and the performance of the observer could be verified using the measured data . i. vladimirov and i. r. petersen , `` coherent quantum filtering for physically realizable linear quantum plants , '' in _ proceedings of the 2013 european control conference _ , zurich , switzerland , july 2013 , arxiv:1301.3154 . z. miao , l. a. d. espinosa , i. r. petersen , v. ugrinovskii , and m. r. james , `` coherent quantum observers for finite level quantum systems , '' in _ australian control conference _ , perth , australia , november 2013 . i. r. petersen , `` a direct coupling coherent quantum observer , '' in _ proceedings of the 2014 ieee multi - conference on systems and control _ , antibes , france , october 2014 , also available arxiv 1408.0399 . , `` a direct coupling coherent quantum observer for a single qubit finite level quantum system , '' in _ proceedings of 2014 australian control conference _ , canberra , australia , november 2014 , also arxiv 1409.2594 . , `` time averaged consensus in a direct coupled coherent quantum observer network for a single qubit finite level quantum system , '' in _ proceedings of the 10th asian control conference 2015 _ , kota kinabalu , malaysia , may 2015 . i. r. petersen and e. h. huntington , `` a possible implementation of a direct coupling coherent quantum observer , '' in _ proceedings of 2015 australian control conference _ , gold coast , australia , november 2015 . m. r. james , h. i. nurdin , and i. r. petersen , `` @xmath153 control of linear quantum stochastic systems , '' _ ieee transactions on automatic control _ , vol . 53 , no . 8 , pp . 17871803 , 2008 , arxiv : quant - ph/0703150 . a. j. shaiju and i. r. petersen , `` a frequency domain condition for the physical realizability of linear quantum systems , '' _ ieee transactions on automatic control _ , vol . 57 , no . 8 , pp . 2033 2044 , 2012 .

this paper considers the problem of constructing a direct coupling quantum observer for a quantum harmonic oscillator system . the proposed observer is shown to be able to estimate one but not both of the plant variables and produces a measureable output using homodyne detection .

gromov s non - squeezing theorem @xcite identified a new rigidity phenomenon in symplectic geometry : the standard symplectic ball can not be symplectically embedded ( symplectically squeezed ) into any cylinder of smaller radius . by contrast , in the contact setting any ( darboux ) ball can be contact embedded into an arbitrarily small neighboourhood of a point . eliashberg - kim - polterovich @xcite therefore studied , in the contact manifold @xmath5 , prequantized balls @xmath11 and a more restrictive notion of squeezing : embedding via a globally defined compactly supported contactomorphism , @xmath12 can be embedded into an arbitrarily small neighborhood of a point by a contactomorphism but in this case the embedding , while smoothly isotopic to the identity , is not compactly supported nor contact isotopic to the identity ; see @xcite . ] which can further be required to be isotopic to the identity in this class . in their terminology , a _ ( contact ) squeezing _ of an open set @xmath13 into an open set @xmath14 is a compactly supported contact isotopy @xmath15}$ ] such that @xmath16 . constructed squeezings in @xcite are of this kind . in their proofs of non - existence of a squeezing , however , they prove a formally stronger statement , namely , non - existence of a compactly supported contactomorphism @xmath17 ( possibly not isotopic to the identity ) such that @xmath18 ; we will call such a contactomorphism a _ coarse squeezing_. it is not known in @xmath5 if this formally stronger non - squeezing statement is strictly stronger . eliashberg - kim - polterovich @xcite showed contact squeezing is closely related to the concept of orderabilitygiven a liouville domain @xmath19 with ideal contact boundary @xmath20 ( see section 1.5 of @xcite for terminology ) , non - squeezing at arbitrarily small scales for fiberwise star - shaped domains " in @xmath21 implies orderability of @xmath20 . ] of a contact manifold introduced by eliashberg - polterovich in @xcite and moreover , squeezing may be possible at one scale and not at another in the same manifold . more precisely , consider @xmath5 with contact structure @xmath22 and liouville form @xmath23 on @xmath24 , @xmath25 , @xmath26 . let @xmath27 denote the identity component of the group @xmath9 of compactly supported contactomorphisms . these are time-1 maps of compactly supported contact isotopies , so squeezing of @xmath13 into @xmath14 is equivalent to the existence of @xmath28 such that @xmath29 . for any positive @xmath30 , let @xmath31 be the ball of symplectic capacity @xmath30 , and @xmath32 its prequantization . eliashberg - kim - polterovich proved : [ thm : ekp ] let @xmath33 then there is a contact squeezing of @xmath7 into itself . by contrast , if @xmath3 is an integer , there is no coarse contact squeezing of @xmath7 into itself . it remained unknown for some time whether non - squeezing in this setting held also for non - integer @xmath34 . in 2010 tamarkin @xcite sketched a proof of the affirmative answer . this was recently proved by chiu : [ thm : main ] let @xmath3 . then there is no contact squeezing of @xmath7 into itself . discussions with tamarkin @xcite also inspired our use of @xmath0-equivariance : we prove theorem [ thm : main ] by means of a @xmath0-equivariant version of the cylindrical contact homology of eliashberg - kim - polterovich @xcite . in fact , as in @xcite , we prove a formally stronger statement : [ thm : main][thm : main - strong ] let @xmath3 . then there is no coarse contact squeezing of @xmath7 into itself . in section [ sec : room ] , we observe that theorem [ thm : main ] also has consequences for squeezings of @xmath7 into itself when @xmath35 : in some cases they require a larger domain of support , i.e. , _ squeezing room _ , than previously established . this is stated in theorem [ thm : main2 ] . [ rem : into - itself ] squeezing of @xmath7 into itself is an open condition . indeed , existence of a squeezing of @xmath7 into itself implies existence of a squeezing of some larger @xmath36 into a smaller @xmath37 and this will in particular squeeze all intermediate prequantized balls into themselves . an equivalent and [ thm : main] refer to existence of a squeezing only ; the required squeezing room could in principle be greater the farther apart @xmath38 and @xmath39 are taken . ] formulation of theorem [ thm : main ] ( resp . [ thm : main] ) is therefore : let @xmath40 , then there is no squeezing ( resp . coarse squeezing ) of @xmath36 into @xmath37 . note the strict inequality @xmath41 which can without loss of generality be imposed . we prove theorem [ thm : main - strong ] by proving an alternate , equivalent statement , theorem [ thm : cover ] . to formulate this , let @xmath42 denote the identity component of the group @xmath43 of compactly supported @xmath0-equivariant contactomorphisms for the @xmath0-action generated by a vertical shift @xmath44 . by analogy with the definition of contact squeezing @xcite , define a _ @xmath0-equivariant ( contact ) squeezing _ of an open set @xmath13 into an open set @xmath14 as @xmath45 such that @xmath46 . define a _ coarse @xmath0-equivariant squeezing _ analogously but requiring only @xmath47 . our main result is : [ thm : cover ] for any prime @xmath48 and any @xmath49 such that @xmath50 , if @xmath51 , there is no @xmath0-equivariant contact squeezing of @xmath36 into @xmath37 , not even a coarse one . [ rem : kgreater2 ] note that case @xmath52 of theorem [ thm : cover ] is implied by the non - existence of a squeezing of @xmath53 into itself ( proved by @xcite , see theorem [ thm : ekp ] above ) . it will therefore be sufficient for the purposes of proving theorem [ thm : cover ] to establish the case @xmath54 and we make this assumption in the computations of section [ sec : groups - preqb ] . the equivalence of theorems [ thm : main - strong ] and [ thm : cover ] follows from properties of the contact @xmath55-fold cover of @xmath56 . indeed , let @xmath57 and , as a manifold , define the @xmath55-fold cover of @xmath56 by the covering map @xmath58 assuming the standard contact structure on both base and cover , one has that @xmath59 is a contactomorphism . deck transformations in the cover are then also contactomorphisms ; they form a cyclic group isomorphic to @xmath0 , generated by @xmath60 . we have : [ lem : equiv ] @xmath61 where @xmath62 is the unique lift ( as a compactly supported diffeomorphism ) of @xmath63 . likewise we have @xmath64 . moreover , for any @xmath65 , @xmath66 . each @xmath67 has a unique lift where both cover @xmath68 and base @xmath69 are @xmath5 and @xmath70 is a compactly supported diffeomorphism . lift @xmath63 to the local diffeomorphism @xmath71 given by @xmath72 . because @xmath73 and @xmath63 has compact support , @xmath63 acts trivially on @xmath74 so @xmath75 maps @xmath76 to the same subgroup of @xmath74 as does @xmath59 , namely @xmath77 and , hence , by the unique lifting property , lifts to @xmath78 such that @xmath79 which is unique up to composition with deck transformations ; of such lifts there is a unique one with compact support . ] to a compactly supported diffeomorphism @xmath62 of the @xmath55-fold cover and this @xmath62 is a contactomorphism since @xmath63 is and the contact structure in the cover is the pullback by @xmath59 of its counterpart in the base . moreover , @xmath62 is @xmath0-equivariant by construction since @xmath60 is a deck transformation . thus @xmath80 . on the other hand any element @xmath81 descends to a well - defined contactomorphism @xmath63 of the base since it commutes with @xmath60 , and so we have @xmath82 . applying this correspondence to isotopies we obtain the statement regarding @xmath42 and @xmath27 . the final statement of the lemma is immediate because @xmath83 . as a corollary , we obtain the claimed equivalence of theorems [ thm : main - strong ] and [ thm : cover ] : [ cor : equivalent ] @xmath7 can be squeezed into itself by @xmath67 if and only if @xmath84 can be squeezed into itself by @xmath80 ( and likewise for @xmath27 and @xmath42 ) . thus , theorem [ thm : main - strong ] is equivalent to theorem [ thm : cover ] . the first statement is immediate from lemma [ lem : equiv ] . for the equivalence of theorems note that , once @xmath55 prime and @xmath50 are fixed , @xmath39 in the hypotheses of theorem [ thm : cover ] can without loss of generality be specified to be arbitrarily close to @xmath85 , in particular such that @xmath86 since squeezing @xmath36 into a smaller @xmath37 would imply squeezing @xmath36 into all @xmath7 for @xmath87 therefore , theorem [ thm : cover ] is equivalent ( putting @xmath88 ) to non - existence - under the hypothesis @xmath55 prime , @xmath50 , and @xmath89 - of @xmath67 which squeezes @xmath90 into @xmath91 . since for every pair @xmath92 there exist and consider intervals @xmath93 , @xmath49 . since @xmath94 , for sufficiently large @xmath95 each @xmath96 overlaps with @xmath97 and so by the infinitude of primes @xmath98 such that @xmath96 contains a prime @xmath55 ; i.e. , @xmath99 . ] @xmath100 with @xmath55 prime such that @xmath101 the above can be restated as non - existence of a squeezing of @xmath90 into @xmath91 when @xmath102 . by remark [ rem : into - itself ] this is equivalent to theorem [ thm : main - strong ] . eliashberg - kim - polterovich @xcite use ( non - equivariant ) cylindrical contact homology to prove that @xmath53 can not be squeezed into itself . they then conclude by a covering space argument that @xmath103 can not be squeezed into itself for any @xmath104 . their two - part proof does not emphasize the inherent @xmath105-action ; however , in the language of the present section , it amounts to first showing @xmath53 can not be squeezed into itself , noting this implies @xmath53 can not be squeezed into itself @xmath105-equivariantly , and then using the last statement of lemma [ lem : equiv ] to conclude @xmath103 can not be squeezed into itself . for the prequantized ball @xmath53 , both @xmath0-equivariant squeezing and ( non - equivariant ) squeezing are impossible and , moreover , @xmath53 is the only prequantized ball for which the contact homology of @xcite can directly - graded vector spaces @xmath106 are isomorphic for all @xmath107 and differ from @xmath108 for @xmath109 . this makes the invariants suitable to directly rule out only squeezing of @xmath7 into @xmath110 ( since we allow squeezings to have arbitrarily large support - see remark [ rem : comparens ] ) . ] rule out squeezing . a priori , however , non - existence of a @xmath105-equivariant squeezing is a weaker notion than non - existence of a squeezing and could potentially be true in more situations . indeed , this is the contribution of the present paper . for @xmath55 prime , we use @xmath0-equivariant contact homology to rule out @xmath0-equivariant squeezing of any @xmath111 , @xmath49 into itself when @xmath50 ( theorem [ thm : cover ] ) , although such prequantized balls are known by @xcite to be squeezable into themselves non - equivariantly ( c.f . theorem [ thm : ekp ] ) . we now define @xmath113 , a @xmath0-equivariant analog of the ( non - equivariant ) cylindrical contact homology @xmath114 developed by eliashberg - kim - polterovich @xcite and kim @xcite for @xmath5 , which also has similarities to a @xmath0-equivariant version of @xmath114 developed by milin @xcite for a different @xmath0-action . like these theories , it lies within the general framework of symplectic field theory proposed by eliashberg - givental - hofer @xcite . we give the construction of @xmath113 as we recall that of @xmath114 , and we explain how well - definedness of @xmath113 follows with only minor modifications from corresponding arguments for @xmath114 due to @xcite , @xcite , @xcite , @xcite with , in addition , the construction of coherent orientations from bourgeois - mohnke @xcite ; these aspects are highlighted in boldface in the construction and addressed in the paragraph technical arguments " . foundational issues with cylindrical contact homology which arise in the presence of multiply covered orbits ( see for example @xcite for a discussion ) are avoided in both our setting and that of eliashberg - kim - polterovich @xcite by taking as generators only closed reeb orbits in the free homotopy class @xmath115 $ ] . in section [ sec : results - preqb ] we state our main results for @xmath116 , showing how they imply non - squeezing . in section [ sec : groups - preqb ] we give proofs of these statements as well as a proof of analogous statements for @xmath117 with @xmath0-coefficients . our aim is for the present paper to serve as an extension of eliashberg - kim - polterovich @xcite in which the reader familiar with @xcite can view non - squeezing at large scale from the vantage point provided by contact homology . [ rem : z2vszk ] to keep the presentation compact : * we assume @xmath118 from now on*. our construction applies equally well to the case @xmath52 but this would require specification of a different projective resolution to compute equivariant homology ( see remark [ rem : equivhom ] ) and result in slightly different must be replaced with multiplication by @xmath119 . ] chain complexes in equivariant computations . since we do not need the case @xmath52 ( see remark [ rem : kgreater2 ] ) we omit it . we restrict attention to @xmath120 with contact structure @xmath22 and @xmath0-action as defined in sections [ sec : intro ] and [ sec : zk - action ] . our construction , however , like that of @xcite , goes through verbatim in prequantizations @xmath21 for many other liouville manifolds @xmath121 ( c.f . theorem 4.47 of @xcite ) . an open domain @xmath122 with compact closure is said to be _ fiberwise star - shaped _ if its boundary @xmath123 is transverse to the fibers @xmath124 , @xmath26 and intersects them along hypersurfaces transverse to the liouville vector field @xmath125 determined by @xmath126 . in particular prequantized balls @xmath7 are fiberwise star - shaped . let @xmath127 , resp . @xmath128 , be the class of domains @xmath129 such that @xmath130 is fiberwise star - shaped , resp . fiberwise star - shaped and @xmath0-invariant , and @xmath131 , resp . @xmath132 . given @xmath133 , we construct the @xmath134-graded vector space @xmath135 , as @xcite constructed @xmath136 for @xmath137 , so that the resulting association is functorial in an invariant way ( invariant under the action of @xmath138 resp . @xmath139 - see theorem [ thm : g - functor ] below , resp . theorem 4.47 of @xcite ) . denote by @xmath140 the set of all _ admissible _ contact forms on @xmath130 , namely forms @xmath141 where @xmath142 is a positive hamiltonian equal to a constant @xmath143 outside a compact set , and the reeb flow of @xmath144 has no contractible closed orbits @xmath145 of action ( i.e. period ) @xmath146 less than or equal to @xmath143 . let @xmath147 consist of those forms which do not have @xmath148 as _ critical value _ , i.e. have no closed reeb orbit @xmath145 of action @xmath149 . denote by @xmath150 those @xmath151 which are @xmath0-invariant . as in @xcite , endow these spaces of contact forms with the anti - natural " partial order @xmath152 : @xmath153 an admissible contact form @xmath144 equal to @xmath154 outside a compact set is said to be _ regular _ if all orbits @xmath145 in @xmath155 = [ \text{pt}\times s^1]\}\ ] ] are _ non - degenerate_. here @xmath156 $ ] denotes the free homotopy class of @xmath145 as a loop @xmath157 and non - degeneracy means the restriction of the poincar return map for the reeb flow of @xmath144 along @xmath145 to the contact hyperplane bundle does not have 1 as eigenvalue . by standard arguments generic admissible contact forms are regular . for regular @xmath144 , @xmath158 is graded by the conley - zehnder index @xmath159 defined in terms of paths of symplectic matrices as in @xcite . note this convention differs by @xmath160 from that in @xcite ( c.f . erratum to @xcite ) . if @xmath144 is @xmath0-invariant , then @xmath161 generates a @xmath0-action on @xmath158 which preserves @xmath162 . using @xmath144 , identify the symplectization @xmath163 of @xmath164 with @xmath165 and put @xmath166 . let @xmath167 and @xmath59 denote the hyperplane bundles on @xmath168 resp . @xmath163 which are respectively pull - backs of the standard contact structure on @xmath169 and @xmath170 on @xmath164 . lift also the @xmath0-action of @xmath164 to @xmath163 , @xmath168 . write @xmath171 for the reeb vector field of @xmath144 . consider almost complex structures @xmath172 on @xmath173 which are _ adjusted _ to @xmath144 in the sense of @xcite : @xmath172 is invariant under translations in the @xmath174-coordinate , @xmath175 and @xmath176 is compatible with @xmath177 , @xmath178 , and finally , @xmath179 outside the symplectization of a compact subset of @xmath164 . when @xmath144 is @xmath0-invariant , a @xmath172 adjusted to @xmath144 which is also @xmath0-invariant is said to be _ @xmath0-adjusted _ to @xmath144 . given an almost complex structure @xmath172 adjusted , resp . @xmath0-adjusted , to a regular admissible @xmath144 , let @xmath180 , resp . @xmath181 , be the vector space generated over @xmath0 by orbits in @xmath158 and @xmath134-graded by @xmath162 . for @xmath182 which are not critical values of @xmath144 , using the action filtration let @xmath183 and define @xmath184 analogously . to define a differential @xmath185 we make no changes to the definition of @xcite , however * transversality * arguments needed to establish well - definedness of @xmath185 for generic @xmath172 require a slight modification as explained below . note : the definition of the differential in @xcite is given in more generality , as part of the construction of generalized floer homology ; for a more accessible , @xmath186-specific description the reader is referred to @xcite . roughly speaking , for both equivariant and non - equivariant theories we consider for closed reeb orbits @xmath187 the moduli space @xmath188 consisting of @xmath172-holomorphic cylinders @xmath189 asymptotic at the ends to @xmath187 such that @xmath190 and @xmath191 , for @xmath192 the periods of @xmath187 respectively , and assuming standard complex structure @xmath193 on @xmath194 and almost complex structure @xmath172 adjusted to @xmath144 on @xmath163 . in the non - equivariant case , for generic almost complex structures adjusted to @xmath144 the moduli space @xmath188 forms a smooth oriented manifold . in the equivariant case , we allow only @xmath0-adjusted complex structures and modify the usual genericity argument as explained below . this manifold @xmath188 , in both equivariant and non - equivariant cases , is acted on by the group @xmath195 of holomorphic re - parametrizations ( of target , domain ) and quotienting by these one obtains a smooth oriented manifold @xmath196 of dimension @xmath197 . since @xmath172 and @xmath144 are standard at infinity and @xmath144 has no contractible closed reeb orbits with action @xmath198 , * compactness * results of @xcite show that @xmath196 compactifies to a moduli space of broken " @xmath172-holomorphic cylinders ; this applies in both the equivariant and non - equivariant frameworks . in the case @xmath199 this moduli space is a compact , oriented @xmath200-dimensional manifold and so consists of a finite number of points with sign . for this , one needs to have arranged a system of * coherent orientations * on moduli spaces in a way which is compatible with glueing , and , in the case of our @xmath0-equivariant theory , invariant under the action of @xmath0 . this extra property comes for free from the construction of @xcite assuming the contact forms and almost complex structures used are @xmath0-invariant ( see below ) . the differential on @xmath180 , resp . @xmath181 is then defined by counting , modulo @xmath55 , all points in @xmath196 with signs . by construction of the compactification @xcite it follows that @xmath201 so @xmath202 , resp . @xmath203 , is a chain complex . moreover , since @xmath204 if and only if @xmath205 and signs of elements of @xmath196 are preserved by @xmath60 , it follows that @xmath203 comes equipped with a @xmath0-action , i.e. , @xmath185 is @xmath0-equivariant for the @xmath0-action induced on spaces of reeb orbits by @xmath60 . @xmath206 is defined as the homology of @xmath207 after showing it does not depend on @xmath172 . likewise we define @xmath208 as the @xmath0-equivariant homology of @xmath209 . note that any contactomorphism , resp . @xmath0-equivariant contactomorphism , @xmath17 will set up a 1 - 1 correspondence not only between chain groups for @xmath144 and @xmath210 but also between moduli spaces defined above so there is a chain map , resp . @xmath0-equivariant chain map , @xmath211 from the respective chain complex for @xmath144 to that for @xmath210 which is an isomorphism in all degrees , yielding ( grading - preserving ) isomorphisms : @xmath212 and @xmath213 * ( equivariant homology computation ) * [ rem : equivhom ] in general , given a @xmath0-action on a chain complex @xmath214 , @xmath0-equivariant homology is defined as follows . let @xmath215 \cong { { \mathbb{z}}}_k[t]/(t^k-1)$ ] be the group ring of the group @xmath0 with @xmath0-coefficients . let @xmath0 act on @xmath0 trivially to make @xmath0 into an @xmath216-module . let @xmath217 be any projective resolution of @xmath0 ( as an @xmath216-module ) and tensor @xmath217 with @xmath214 over @xmath216 . the @xmath0-equivariant homology of @xmath214 is defined to be the usual homology of this tensor product . up to isomorphism this is independent of the choice of @xmath217 since all projective resolutions are quasi - isomorphic as @xmath216-chain complexes ( to @xmath218 and so to each other ) and thus the resulting tensor products are also quasi - isomorphic . in computations , we follow milin @xcite and - assuming @xmath118 - use the projective resolution @xmath219 given @xmath220 , @xcite define a monotonicity chain map @xmath221 by considering the moduli space @xmath188 of @xmath172-holomorphic cylinders between @xmath187 which are closed reeb orbits for @xmath222 respectively , where @xmath172 is an almost complex structure adjusted to an admissible concordance structure " between @xmath222 . this construction is then extended to the case @xmath223 , i.e. , @xmath224 . in our equivariant setting we follow the identical procedure using however @xmath0-invariant ingredients . to describe this more precisely , assume @xmath220 where @xmath225 , @xmath226 . let @xmath227 and put @xmath228 $ ] with coordinates @xmath229 . consider any function @xmath230 such that @xmath231 , and , outside a compact subset of @xmath164 , @xmath142 depends only on @xmath232 , while near the respective boundaries , @xmath233 , @xmath142 is linear in @xmath232 of the form @xmath234 . in particular , using such an @xmath142 , @xmath236 can be identified with the region @xmath237 , @xmath238 via the map @xmath239 . view @xmath240 as the symplectization of @xmath164 with symplectic form @xmath241 and denote the pull - back of this 2-form to @xmath242 as @xmath243 . the pair @xmath244 is said to be an _ ( admissible ) concordance structure _ between @xmath222 , or in the equivariant setting an _ ( admissible ) @xmath0-concordance structure _ if @xmath142 is also @xmath0-invariant ( for the lifted @xmath0-action ) . the linearity of @xmath142 in @xmath232 near the boundaries , @xmath233 , of @xmath236 ( c.f . normal form " in @xcite ) gives neighbourhoods of these boundaries already locally the structure of symplectizations of @xmath245 respectively . by extending @xmath142 linearly to all of @xmath174 and extending @xmath246 accordingly , @xmath247 acquires the structure of a symplectic cobordism between @xmath248 and @xmath249 . moreover , when @xmath142 is @xmath0-invariant so is the symplectic structure @xmath246 on @xmath250 . warning : though @xmath251 , we use different names to recall the different symplectic structures - @xmath252 is a cobordism between @xmath245 while @xmath253 is the symplectization of @xmath254 . an almost complex structure @xmath172 on @xmath236 is _ adjusted _ to the concordance @xmath255 between @xmath222 for specific choices , @xmath256 , of respectively adjusted almost complex structures on symplectizations @xmath257 of @xmath245 if @xmath246 tames @xmath172 , @xmath172 is pseudoconvex at infinity and @xmath172 agrees with @xmath256 near the boundaries @xmath233 . the second condition means that @xmath258 is foliated by weakly @xmath172-convex hypersurfaces outside the symplectization of a compact subset of @xmath164 . in the equivariant setting , an adjusted @xmath172 which is also @xmath0-invariant is said to be _ @xmath0-adjusted_. in either case , we denote also by @xmath172 the extension of @xmath172 to @xmath250 . analogous to when we defined the differential @xmath185 , given @xmath220 we consider in both equivariant and non - equivariant theories the moduli space @xmath188 consisting of @xmath172-holomorphic cylinders @xmath259 asymptotic at the ends to @xmath187 such that @xmath190 and @xmath191 , for @xmath192 the periods of @xmath187 respectively , and assuming standard complex structure @xmath193 on @xmath194 and almost complex structure @xmath172 adjusted to a chosen concordance between @xmath222 . in the non - equivariant setting standard transversality arguments ( as in @xcite , @xcite ) establish that for generic @xmath172 the space @xmath188 is a smooth oriented manifold ( assuming once again a system of coherent orientations ) . slight modifications of these arguments ( see below ) apply in our equivariant setting implying the same statement for generic almost complex structures @xmath0-adjusted to a @xmath0-concordance . the manifold @xmath188 is acted upon freely by the re - parametrization in the @xmath232-direction ) ] group @xmath194 and quotienting yields a smooth manifold @xmath196 of dimension @xmath260 , which compactifies to a moduli space of broken @xmath172-holomorphic cylinders by @xcite . when @xmath261 the space @xmath196 is a finite collection of points with sign which we count modulo @xmath55 and the chain map @xmath262 is defined , exactly as in @xcite , by setting @xmath263 to be the sum over all @xmath264 weighted by this count @xmath265 . this induces the ( grading - preserving ) monotonicity morphism @xmath266 . three natural grading - preserving morphisms , besides @xmath267 , are important - those due to scaling invariance , contactomorphism invariance ( @xmath211 ) and window enlargement . we ve given @xmath211 . the other two are also immediate in both equivariant and non - equivariant settings from the definitions . assume @xmath268 , then there are chain maps @xmath269 where the second map is an isomorphism in all gradings under the additional hypothesis that @xmath144 has no closed reeb orbits with action in the window @xmath270 $ ] , and @xmath271 denotes the re - scaled version of @xmath172 adjusted to @xmath272 . in general , the composition of @xmath273 and @xmath274 , in either order , gives @xmath267 ( see remark 4.40 in @xcite which applies verbatim in our setting ) . in particular , this allows to extend the definition of @xmath267 to the case @xmath275 by first scaling @xmath276 by a suitably small factor @xmath277 . moreover , by definition , concordances behave well under compactly supported contactomorphisms @xmath17 since these preserve the contact forms at infinity and thus induce a bijective correspondence between concordances . indeed , given contact forms @xmath278 and induced forms @xmath279 , let @xmath230 be a function defining a concordance structure between @xmath222 . then the function @xmath280 , @xmath281 defines a concordance structure between @xmath282 where @xmath283 such that @xmath284 and @xmath285 is induced from @xmath246 by @xmath286 . this implies that @xmath267 commutes with @xmath211 on chain level . the identical argument applies in the equivariant setting assuming @xmath132 and @xmath222 is @xmath0-admissible . passing to homology , the morphism @xmath224 induces a morphism @xmath287 which commutes with @xmath211 , and likewise in the equivariant case , for @xmath17 in the respective group . in principle @xmath267 still depends on both the choice of concordance @xmath142 and of adjusted almost complex structure @xmath172 , while @xmath288 , @xmath289 and @xmath274 depend on @xmath172 . by a straightforward argument ( see proposition 4.30 of @xcite ) which goes through verbatim in our setting the vector spaces @xmath288 , @xmath289 are independent of @xmath172 . finally , by standard floer theoretic arguments applied to a homotopy between concordances or between almost complex structures adjusted to a concordance ( see page 1692 of @xcite ) , it follows that monotonicity morphisms @xmath267 are independent of the choices of @xmath142 and @xmath172 . we omit @xmath172 in the notation from now on . let @xmath133 . the morphisms @xmath267 make the family of vector spaces @xmath290 into a directed system over @xmath291 and after taking its direct limit , the morphisms @xmath274 make resulting vector spaces @xmath292 into an inverse system over @xmath293 . following @xcite we define @xmath294 and make the same definition without superscripts @xmath0 for @xmath137 . then given @xmath295 there is an induced morphism @xmath296 and the properties of and @xmath211 pass to the double limit yielding the following statement in terms of @xmath297-functors acting on category @xmath127 , a @xmath297-functor on @xmath127 is a functor @xmath142 and a family of natural transformations @xmath298 , @xmath299 , such that @xmath300 for all @xmath301 ; see also the reference to jackowski and slominska in @xcite ] , stated for the non - equivariant setting on p. 1650 of @xcite ( see also their theorem 4.47 ) : [ thm : g - functor ] @xmath302 is a @xmath297-functor from @xmath128 to @xmath134-graded vector spaces , for @xmath303 . whereas eliashberg - kim - polterovich @xcite restricted to @xmath304-coefficients in order to streamline their presentation avoiding the need for coherent orientations , these can be produced by the now - standard construction of bourgeois - mohnke @xcite ( see also bourgeois - oancea @xcite ) which pulls back orientations from the orientations of determinant line bundles over certain spaces of fredholm operators . when all ingredients in the construction are @xmath0-invariant , the orientations on these bundles are as well , hence so too are their pull - backs . this gives coherent orientations in our equivariant setting and also means @xmath114 of @xcite is well - defined with @xmath0-coefficients ; we re - prove their results for @xmath7 using @xmath0-coefficients , @xmath118 in the next section . the compactification results used to define the differential in @xcite are obtained as a consequence of the restriction on the action of closed contractible reeb orbits for admissible contact forms and the requirement that almost complex structures be standard outside the symplectization of a compact subset @xmath305 of @xmath164 . this second condition means @xmath306 is foliated by weakly @xmath172-convex hypersurfaces so all @xmath172-holomorphic cylinders project to @xmath305 . given this and the absence of contractible closed reeb orbits in @xmath158 , the results from @xcite imply the needed compactification and glueing statements proving @xmath307 . the above goes through verbatim in our equivariant setting as well . finally , we address transversality . because the @xmath0-action we consider is a covering space action , only minor modifications of the standard arguments for the non - equivariant setting of @xcite are required . by contrast , milin s @xmath0-equivariant theory requires more intricate arguments because her @xmath0-action , generated by @xmath308 , is not free ( it fixes all points of @xmath309 ) . the standard transversality arguments ( see @xcite for a detailed account ) can be summarized as follows . to show the moduli space @xmath310 which depends on a specific @xmath172 is , for generic @xmath172 , a banach manifold of specified dimension , one considers , in the setting of @xcite , the universal moduli space @xmath311 where @xmath312 is the smooth banach manifold of almost complex structures adjusted to @xmath144 , resp . adjusted to a concordance . this space , @xmath311 , can be identified with the zero set of the cauchy - riemann operator @xmath313 defined on the product @xmath314 , where @xmath315 is the smooth banach manifold consisting of cylinder maps satisfying all conditions for elements of @xmath188 except the cauchy - riemann equation . by an infinite dimensional implicit function theorem , it suffices to show that the linearized operator @xmath316 is surjective for all @xmath317 in @xmath311 , then it follows that @xmath311 is a banach manifold and so @xmath318 is as well for all regular values @xmath172 of the projection map @xmath319 , while , by sard - smale , the regular @xmath172 are generic . in our @xmath0-equivariant setting we consider instead analogous spaces @xmath320 consisting of @xmath0-adjusted almost complex structures . the only part of the above argument we must modify is the proof that @xmath316 is surjective for all @xmath317 since in the @xmath0-equivariant setting this operator is now defined on the smaller space @xmath321 where @xmath320 consists of @xmath0-adjusted almost complex structures . in fact surjectivity here follows from that on @xmath314 by a covering space argument for @xmath320 as a @xmath55-fold cover of @xmath312 but we instead describe how the surjectivity argument on @xmath314 can be directly modified . surjectivity in the specific case where @xmath322 and @xmath323 is a vertical cylinder is immediate , in equivariant as well as non - equivariant settings , since @xmath316 decomposes as a direct sum of surjections ( see @xcite ) . for other @xmath323 , in the non - equivariant setting surjectivity of @xmath316 is obtained by showing that for any @xmath324 , every non - trivial , finite energy @xmath172-holomorphic cylinder @xmath325 in @xmath163 resp . @xmath250 has an _ injective point _ @xcite , @xcite . this is a point @xmath327 such that @xmath328 and @xmath329 . for genericity of almost complex structures @xmath172 adjusted to concordances a similar statement holds , and likewise for homotopies @xmath330 , @xmath331 $ ] of almost complex structures adjusted to a single concordance ( one can perturb to a homotopy of regular structures , fixing endpoints ) . in all three cases , the existence of needed injective points is guaranteed by the fact that near the ends of each holomorphic cylinder @xmath332 restricts to an embedding @xcite . one injective point implies a neighbourhood of such in every @xmath172-holomorphic cylinder @xmath323 and this forces @xmath316 to be surjective . in all three cases , to pass from existence of an injective point @xmath327 on every cylinder @xmath323 to surjectivity of @xmath316 for all @xmath317 the key ingredient is that one can choose a tangent vector @xmath333 supported in a small ball around @xmath334 with complete freedom due to injectivity of @xmath332 at @xmath327 . this allows to derive a contradiction should @xmath335 fail to be surjective . in our @xmath0-equivariant framework where we work with @xmath320 , in order to carry this argument out it suffices to find a stronger kind of injective point , one which also satisfies @xmath336 . we call such a point _ @xmath0-injective _ avoid @xmath309 , a consideration not needed in our case ] . its existence means a tangent vector @xmath337 with support in @xmath338 can still be chosen with complete freedom given a small ball @xmath69 at @xmath327 and so the usual argument ( see @xcite ) goes through verbatim . first we remark that when @xmath323 is @xmath0-equivariant for the @xmath0-action on @xmath194 generated by a @xmath10-shift in the @xmath339 factor , then it is the lift to the @xmath55-fold cover of a @xmath340-holomorphic cylinder @xmath341 from the base of @xmath194 to the non-@xmath0-invariant base @xmath342 of @xmath258 resp . @xmath163 where @xmath340 is the non-@xmath0-invariant almost complex structure in the base which lifts to @xmath172 . we know @xmath343 is surjective , so @xmath316 is as well . we then consider @xmath323 which are not @xmath0-equivariant . in this case , we claim the image of @xmath323 can not be @xmath0-invariant . indeed if it were then @xmath323 and @xmath344 would be related by a holomorphic re - parametrization of @xmath345 , but this is necessarily a translation and amounts asymptotically along @xmath187 to a shift by @xmath10 , hence is exactly a shift by @xmath10 , i.e. @xmath323 is @xmath0-equivariant . now , because @xmath323 is not @xmath0-invariant we know @xmath346 for any @xmath347 . at the same time , intersection points of holomorphic curves can accumulate only at critical values of both curves . we take a point @xmath327 which is injective for @xmath323 in the usual sense and consider a neighbourhood @xmath130 of @xmath327 consisting also of injective points ( by openness of this condition ) . since @xmath348 can intersect the other holomorphic cylinders @xmath349 , @xmath347 at only finitely many points , there is necessarily a point @xmath350 ( and hence small neighbourhood @xmath351 ) which is @xmath0-injective . in this section we state two results , theorems [ thm : comp - equiv ] and [ thm : psi - perturbed ] , concerning @xmath0-equivariant contact homology of prequantized balls and show that together they imply our main result , theorem [ thm : cover ] . for comparison , we also state theorem [ thm : comp - nonequiv ] , a non - equivariant version of theorem [ thm : comp - equiv ] which is a direct analog with @xmath0-coefficients of eliashberg - kim - polterovich s result for @xmath106 with @xmath304-coef - ficients ( theorem 1.28 and page 1721 of @xcite , with different grading convention ) . proofs of these results are then given in section [ sec : groups - preqb ] . we remark that our proof to simplify the presentation of @xmath0-equivariant proofs ( see remark [ rem : z2vszk ] ) this non - equivariant argument goes through without change for @xmath304-coefficients . ] of theorem [ thm : comp - nonequiv ] uses only contact homology , and does not pass to generalized floer homology as did the proof of eliashberg - kim - polterovich ; it thus provides an alternative , more direct , proof of their result as well . we write @xmath352 $ ] for the integer part of @xmath353 . [ thm : comp - nonequiv ] when @xmath354 @xmath355 \\ 0 , & \text { otherwise . } \end{array } \right .\ ] ] moreover , if @xmath356 = [ 1/r_2]$ ] then the morphism induced by inclusion @xmath357 is an isomorphism in the grading @xmath358=-n-2n[1/r_2]$ ] . by comparison , for @xmath0-equivariant contact homology we have : [ thm : comp - equiv ] when @xmath359 and @xmath354 @xmath360 \\ 0 , & \text { if } m < -n -2n[1/r ] . \end{array } \right .\ ] ] moreover , if @xmath361 then the morphism induced by inclusion @xmath357 is an isomorphism in all gradings @xmath362 $ ] . broadly speaking , the basic computations for @xmath106 and @xmath363 are similar in that - in both cases - we find cofinal sequences of contact forms @xmath144 such that the chain complex @xmath364 is quasi - isomorphic to @xmath365 with non - trivial chain module in degree @xmath366 $ ] . a major difference in the two cases , however , is that non - equivariant homology @xmath367 is non - trivial iff @xmath368 , while @xmath0-equivariant homology @xmath369 is non - trivial iff @xmath370 . not only can we conclude from theorem [ thm : comp - equiv ] that the morphism @xmath371 induced by an inclusion @xmath372 is an isomorphism in all degrees for which both @xmath373 and @xmath374 are non - trivial , a stronger , @xmath17-perturbed , result also holds : [ thm : psi - perturbed ] given @xmath375 and @xmath376 such that @xmath377 , the inclusion morphism @xmath378 is an isomorphism in all degrees for which both @xmath379 and @xmath380 are non - trivial . given the @xmath297-functoriality of @xmath381 for @xmath303 ( theorem [ thm : g - functor ] ) , theorems [ thm : comp - equiv ] and [ thm : psi - perturbed ] together imply theorem [ thm : cover ] : suppose @xmath382 squeezes @xmath36 into @xmath37 and @xmath51 . by remark [ rem : into - itself ] we may without loss of generality assume @xmath39 to be as close to @xmath85 as desired , in particular , assume @xmath383 . by hypothesis @xmath384 < \ell < [ \frac 1 { r_2}]$ ] . we have @xmath385 . by theorem [ thm : psi - perturbed ] the inclusion morphism @xmath386 is an isomorphism for @xmath327 in the grading range @xmath387 $ ] . let @xmath388 . then @xmath389 by theorem [ thm : comp - equiv ] ; hence , @xmath390 by theorem [ thm : g - functor ] . this is a contradiction , as the isomorphism @xmath386 must , by theorem [ thm : g - functor ] , factor through @xmath391 . [ rem : comparens ] the proof just given for theorem [ thm : cover ] is in the same general spirit as eliashberg - kim - polterovich s proof that @xmath36 can not be squeezed into @xmath37 for @xmath392 . their argument ( generalized in proposition 1.26 of @xcite ) uses the non - vanishing of an inclusion morphism @xmath393 for large such that @xmath394 contains the support of @xmath17 . ] @xmath395 and @xmath396 both of which are given when @xmath392 . this strategy does not work verbatim in our setting because @xmath397 implies @xmath398 in all gradings @xmath399 . a squeezing @xmath17 of @xmath36 into @xmath37 implies , however , multiple interleavings , @xmath400 , so one is not confined to have the smaller @xmath401 " be the monkey - in - the - middle " : we use the first pair of inclusions , while @xcite use essentially the final pair . as we will see in section [ sec : groups - preqb ] , theorem [ thm : psi - perturbed ] which played a key role in our proof is also very much related to the isomorphism @xmath402 which holds for large @xmath395 . to prove all parts of these theorems ( and later theorem [ thm : psi - perturbed ] as well ) we use similar hamiltonians . we first describe these in detail then give arguments for : ( i ) the first statements of both theorems , ( ii ) monotonicity morphisms when @xmath356=[1/r_2]$ ] , and ( iii ) monotonicity morphisms when @xmath356 < [ 1/r_2]$ ] . let @xmath403 , assume @xmath404 and fix @xmath405 . let @xmath406 $ ] ( so @xmath407 ) and consider a 1-dimensional family and make other simplifications . ] of piecewise - linear functions @xmath408 , @xmath409 as shown in fig . [ fig : onereeb ] such that the value @xmath410 is determined by @xmath411 as specified below , and on @xmath7 whose smoothings define contact forms @xmath412 with a single closed reeb orbit at @xmath309 of action @xmath413 and no other closed reeb orbits with action in the window @xmath414 . here @xmath415 denotes @xmath416 $ ] and we assume @xmath417 so there are no closed reeb orbits corresponding to tangencies at @xmath418 . orbits for @xmath419 on the other hand do not contribute when @xmath148 is small since their action is bounded below by @xmath420.,width=264 ] @xmath142 consists of three linear pieces : a left - most linear piece of slope @xmath421 which passes through @xmath422 and @xmath423 , a middle linear piece of twice that slope which passes through @xmath424 , and a right - most piece , on the interval @xmath425 which is horizontal . the parameter @xmath426 is required to be an increasing function @xmath427 of @xmath411 such that @xmath428 , and @xmath429 for sufficiently large @xmath411 . for example @xmath430 satisfies this inequality for @xmath431 . a quick computation shows that the conditions on the middle segment of @xmath432 , namely its slope and choice of @xmath433 , imply @xmath434 . abusing notation , we define the hamiltonian @xmath432 on @xmath24 to be the above function of the variable @xmath435 , for @xmath436 . this lifts to an @xmath339-invariant hamiltonian on @xmath5 equal to @xmath437 outside @xmath7 ; when clear from context we will use the same name to refer to @xmath432 as a function of the real variable @xmath438 or as a hamiltonian on @xmath24 or @xmath5 . put @xmath439 for suitable @xmath440 such that @xmath441 . then @xmath442 for all @xmath443 since the family @xmath432 is non - decreasing in @xmath411 . we consider the sequence @xmath444 where @xmath445 . among hamiltonians which are constantly equal to @xmath437 outside @xmath7 this is a dominating ( i.e. , cofinal ) set . we will now smooth each @xmath446 in small neighbourhoods of its corners in such a way that the new tangent lines remain sufficiently close " to those of the original @xmath446 and the sequence of functions remains non - decreasing and dominating . to make these conditions precise , assume the smoothing operator @xmath447 has one parameter @xmath448 which determines the fineness of the smoothing such that : ( 1 ) by decreasing @xmath449 , @xmath450 becomes arbitrarily @xmath451-close to @xmath142 , ( 2 ) whenever @xmath142 is linear on @xmath452 then @xmath453 , and ( 3 ) when @xmath452 contains a single corner of @xmath142 , @xmath454 with @xmath455 convex in the case @xmath142 is convex over the interval , and @xmath456 with @xmath455 concave in the case of concativity . consider @xmath457 for some @xmath458 . for sufficiently small @xmath449 , @xmath459 is a positive , non - increasing function with @xmath460 for small @xmath438 and @xmath461 for @xmath462 . it is easily verified the contact form @xmath463 has one closed reeb orbit at @xmath309 of action @xmath464 and spaces of closed reeb orbits of action @xmath420 along spheres of constant @xmath438-value such that @xmath465 for some @xmath466 . we denote each such @xmath438-value by @xmath467 and the sphere @xmath468 by @xmath469 . in this case , @xmath470 . on the other hand , for small @xmath449 , tangent lines to @xmath471 are well approximated by so - called generalized tangent lines " to the original piecewise linear @xmath142 : these are either true tangent lines at points @xmath327 of linearity or else lines of slope @xmath472 $ ] which pass through a corner point @xmath327 where linear portions with slopes @xmath473 meet . as a result , one can read off the closed reeb orbits of @xmath474 and their approximate actions from the graph of the un - smoothed @xmath142 ( c.f . remark 5.2 @xcite ) : @xmath471 has one isolated orbit at @xmath309 with action @xmath413 where @xmath411 is the vertical intercept of @xmath142 , and it has an @xmath169-family of orbits with action arbitrarily close to @xmath420 corresponding to each generalized tangent line to @xmath142 at @xmath419 that is horizontal ( @xmath475 ) or has horizontal intercept @xmath476 ( @xmath477 ) with @xmath478 being in either case the vertical intercept of the generalized tangent line , always less than @xmath479 . if one imposes @xmath480 then ( with @xmath481 ) this implies there are no generalized tangent lines at @xmath418 with horizontal intercept @xmath482 , @xmath2 and , thus , no corresponding closed reeb orbits ( see fig . [ fig : onereeb ] ) . otherwise , in general , there will be reeb orbits corresponding to @xmath418 and all have action greater than @xmath411 ( see fig . [ fig : manyreeb ] ) . let @xmath483 respectively be the @xmath438-values where such tangencies to the smoothing occur , and @xmath484 the corresponding spheres . we now require the @xmath485 which defines a smoothing of @xmath446 , @xmath486 to be sufficiently small that the true vertical intercepts of the tangent lines to the smoothing have the same ordering and lie on the same side of @xmath487 as those of generalized tangent lines to @xmath446 . note we have fixed @xmath488 while computing @xmath489 . in addition @xmath485 should be sufficiently small that the @xmath485-windows centred at corners of @xmath446 do not overlap . finally because @xmath444 is non - decreasing we can take each @xmath485 small enough that the smoothing @xmath490 will satisfy @xmath491 and because @xmath492 we can further reduce @xmath485 if necessary so that @xmath493 . inductively , the first condition implies the sequence @xmath494 is non - decreasing ; the second condition will be used to control monotonicity morphisms . the above produces @xmath495-invariant morse - bott functions @xmath496 on @xmath497 with critical submanifolds @xmath309 and the spheres @xmath469 ( if any ) . we now perturb each @xmath496 following the morse - bott computational framework of bourgeois @xcite in order to obtain a morse function @xmath498 on @xmath24 which is @xmath0-invariant near each @xmath469 ( for a @xmath0-action to be specified ) and for which we know certain parts of the associated contact homology chain complex @xmath499 . the perturbing functions are constructed as follows . choose an arbitrarily @xmath451-small @xmath0-invariant morse function @xmath500 with @xmath55 maxima @xmath501 and @xmath55 minima @xmath502 , @xmath503 and @xmath504 , for the @xmath0-action generated by rotation through @xmath505 . now consider an arbitrarily @xmath451-small morse function @xmath506 with one critical point of each even index @xmath507 , @xmath508 and its pullback @xmath509 via the hopf bundle map @xmath510 . this is a morse - bott function on @xmath169 with critical submanifolds which are isolated hopf circles . perturb @xmath511 by adding to it a radially attenuated extension of @xmath512 supported in a small tubular neighbourhood of each critical hopf circle . this produces a morse function @xmath513 with @xmath55 critical points each of index @xmath507 and @xmath514 for @xmath515 and whose morse complex is 0 & & r & ^p(t ) & r & ^ ( t-1 ) & r & ^p(t ) & & ^ ( t-1 ) & r & ^p(t ) & & ^(t-1 ) & r&&0 where each arrow is multiplication by the specified polynomial @xmath516 or @xmath517 . the function @xmath518 is @xmath0-invariant for any ; we need vary @xmath415 because our @xmath0-action induces on each @xmath469 a different @xmath0-action , namely that generated by multiplication by @xmath519 . ] @xmath0-action on @xmath520 generated by multiplication by @xmath519 , @xmath521 . fix a particular choice of @xmath415 and re - label the @xmath522 critical points in every index if necessary so that permutation of generators of the chain modules @xmath216 induced by the chosen @xmath0-action on @xmath169 is given by multiplication by @xmath523 . then the above chain complex is the @xmath0-equivariant morse complex for @xmath518 for the chosen @xmath0-action on @xmath169 . let @xmath498 denote the morse function on @xmath24 which is obtained by adding to @xmath496 a radially attenuated extension of @xmath518 supported in a small tubular neighbourhood @xmath524 of each critical sphere @xmath469 of @xmath496 . on each @xmath525 there are then @xmath526 non - degenerate closed reeb orbits for @xmath527 which may be identified consistently with the @xmath526 critical points of @xmath498 on @xmath469 . the vertical shift contactomorphism @xmath60 on @xmath5 moreover permutes these orbits in @xmath525 according to multiplication by @xmath519 of the corresponding critical points of @xmath498 in @xmath469 . since @xmath528 < k$ ] , the action is free , i.e. , @xmath521 ( this is important ) . as above , assume labeling of the @xmath55 critical points of each index on @xmath469 such that the associated permutation of these points as generators @xmath529 of the chain module @xmath216 is given by multiplication by @xmath523 . on the small neighbourhood @xmath524 of each @xmath469 , the function @xmath498 is a @xmath0-equivariant morse function for the @xmath0-action given by multiplication by @xmath519 . by @xcite the conley - zehnder indices of closed reeb orbits for the contact form @xmath530 are determined by morse indices of corresponding critical points of @xmath498 and the differentials in the chain complex @xmath531 corresponding to concordances between critical points in a common sphere @xmath469 are the same as the corresponding differentials in the @xmath0-equivariant morse complex of @xmath498 . when @xmath488 is sufficiently small , so that only closed reeb orbits corresponding to @xmath532 and @xmath533 have action in the window @xmath414 , then letting @xmath534 $ ] ( which may be greater than @xmath528 $ ] ) and @xmath535 be such that @xmath536 $ ] , respective conley - zehnder indices of closed reeb orbits for @xmath498 are @xmath537 and the chain complex @xmath538 is therefore @xmath539 \to^{d_{m_0 + 1 } } c[-n-2n(m_0 + 1 ) ] \to \cdots \to c[-n-2n(m_0+\ell ) ] \to 0 $ ] where @xmath540 $ ] denotes a sub - complex which is the morse complex for @xmath518 shifted in grading by @xmath193 ; only the @xmath0-equivariant differentials @xmath541 corresponding to concordances between reeb orbits on adjacent spheres @xmath469 and @xmath542 are not immediately known . we assume once again ( as for the operator @xmath447 ) that the perturbations used to construct @xmath498 are suitably @xmath451-small and carried out successively on @xmath496 , @xmath486 . \(i ) to simplify the notation put @xmath543 . these contact forms , which are @xmath339-invariant hence @xmath0-invariant , constitute a cofinal sequence not only for @xmath544 but also for @xmath545 . moreover , each morse - bott function @xmath496 has only one critical point , namely the origin , and so is already a @xmath0-invariant morse function ; thus , @xmath546 and we retain the name @xmath496 . since @xmath547 has only the closed reeb orbit at @xmath309 and this has conley - zehnder index @xmath160 , for all sufficiently small @xmath404 the chain complex @xmath548 for each @xmath458 is @xmath365 with non - trivial chain module in degree @xmath160 and hence @xmath549 and @xmath550 satisfy the formulae we want for @xmath106 , resp . @xmath363 . it remains to check that monotonicity morphisms in both equivariant and non - equivariant contact homology are isomorphisms for chosen small @xmath404 ; the above result for specific @xmath551 will then pass to the double - limit . we give the argument for the non - equivariant case to fix notation but the equivariant case is identical . fix @xmath458 , and recall that @xmath552 . to avoid excessive subscripts put @xmath553 . on the one hand , there is a window enlargement isomorphism @xmath554 because @xmath555 has no closed reeb orbits with action in the window @xmath556 $ ] ( recall @xmath557 ) . on the other hand , using scaling invariance , we also have an isomorphism @xmath558 . composing these we obtain the monotonicity morphism @xmath559 is an isomorphism . since it factors through @xmath560 and all vector spaces @xmath561 are either trivial or @xmath0 this completes the proof of the first statements in theorems [ thm : comp - nonequiv ] and [ thm : comp - equiv ] . \(ii ) to prove the second statement of theorem [ thm : comp - nonequiv ] we take essentially the same sequence of hamiltonians for both @xmath36 and @xmath37 , however , as functions of their respective canonical coordinates @xmath562 and @xmath563 where @xmath564 . remark that @xmath565 @xmath566 @xmath567 @xmath566 @xmath568 and the condition @xmath569 = [ 1/r_2]$ ] implies @xmath570 so @xmath571 for both @xmath572 . now let @xmath573 but still less than @xmath574 and put @xmath575 , @xmath576 . then @xmath577 for both @xmath578 . define hamiltonians @xmath579 for @xmath36 using @xmath580 , and hamiltonians @xmath581 for @xmath37 using @xmath582 . it follows that @xmath583 and @xmath432 both coincide near @xmath309 with the hamiltonian @xmath584 . on the other hand , assume the same function @xmath332 is chosen to define @xmath433 in both families , @xmath432 and @xmath583 . in the first case , @xmath332 gives values of @xmath585 ; in the second case of @xmath586 . since these are values at which lower corners of the respective hamiltonian occur and @xmath587 , @xmath572 we see that @xmath432 s lower corner has larger @xmath438-value than @xmath583 s and so @xmath588 for all @xmath411 . as in ( i ) we now consider only @xmath443 for suitable @xmath589 such that @xmath590 . this implies @xmath591 and if necessary we adjust takes values between @xmath592 and @xmath437 we can re - parametrize it so that an integer value @xmath593 yields @xmath594 arbitrarily close to @xmath595 as desired . ] the definition of @xmath332 so that at least the minimal @xmath596 satisfies @xmath597 . now consider sequences @xmath598 , @xmath486 and assume smoothings are done as above to yield @xmath495-invariant morse - bott functions @xmath496 and @xmath599 respectively . these each have a single critical point and so are already @xmath0-invariant morse - functions . put @xmath600 and @xmath601 . by the same argument as before , each @xmath602 and each @xmath603 has the desired equivariant and non - equivariant contact homology and these pass to the double limit . we claim that for fixed @xmath404 sufficiently small and @xmath486 sufficiently large the monotonicity morphism induced by @xmath604 is an isomorphism , thus yielding the desired statement in the double limit . indeed , @xmath605 for sufficiently large @xmath606 because @xmath607 is a dominating sequence for @xmath37 and @xmath608 for all @xmath609 with @xmath597 ( recall @xmath610 corresponds to @xmath611 ) . at the same time @xmath612 induces an isomorphism in contact homology for all @xmath486 by the arguments above so by functoriality and the fact that each @xmath613 is either trivial or 1-dimensional , the inclusion morphism @xmath614 is an isomorphism in all gradings @xmath415 , in particular , the grading @xmath615 where both vector spaces are non - trivial . note chain maps @xmath616 induced by the @xmath0-equivariant chain maps @xmath617 are then also isomorphisms in all gradings so , taking the double limit , we obtain moreover the special case @xmath356 = [ 1/r_2]$ ] of the second statement of theorem [ thm : comp - equiv ] . \(iii ) finally , to prove the second statement of theorem [ thm : comp - equiv ] when @xmath356 < [ 1/r_2]$ ] , consider once again hamiltonians as above for @xmath36 using @xmath580 , and for @xmath37 using @xmath582 , where @xmath618 for some choice of @xmath619 . let @xmath620 $ ] , @xmath621 $ ] . as in fig . [ fig : onereeb ] but where we allow @xmath622 , denoting @xmath416 $ ] once again by @xmath415 . smoothings of such @xmath623 define contact forms @xmath624 which have not only the closed reeb orbit @xmath309 with action in the window @xmath414 as in fig . [ fig : onereeb ] , but additionally closed reeb orbits corresponding to generalized tangent lines at @xmath625 ; unlike orbits for tangencies at @xmath626 , these orbits have action in the window @xmath414 . , height=188 ] now , unlike before , we can no longer simultaneously - for @xmath583 and @xmath432 - prevent generalized tangent lines at the upper corners which have horizontal intercept of the form @xmath482 . indeed , taking @xmath627 ( but less than @xmath628 ) and @xmath68 sufficiently large we prevent such lines for @xmath432 as in fig . [ fig : onereeb ] , but then we necessarily have @xmath629 ( because @xmath630 ) so such tangent lines will occur for @xmath583 as in fig . [ fig : manyreeb ] . we can guarantee , however , that their horizontal intercepts @xmath631 occur only for @xmath415 such that @xmath632 by taking @xmath627 and restricting to @xmath443 for some @xmath589 such that @xmath633 . note , for @xmath634 these conditions reduce to those imposed in paragraph ( ii ) . as before , take sufficiently large @xmath68 and now consider perturbed smoothings @xmath635 and @xmath636 which are @xmath0-invariant morse functions . note : unlike in ( ii ) , the functions @xmath599 were morse - bott with critical submanifolds @xmath469 , @xmath632 . note @xmath637 . now , fix @xmath404 and take @xmath486 sufficiently large that @xmath638 . let @xmath639 be a scaling factor such that the line @xmath640 is tangent to and lies below @xmath641 . then @xmath642 and since @xmath643 , the action of all reeb orbits corresponding to tangent lines to @xmath644 at points near the former lower corner " of @xmath641 will be strictly lower that @xmath645 . let @xmath646 be greater than these actions . note that @xmath647 is then an isomorphism ( by composing window enlargement and scaling invariance isomorphisms ) and must factor through @xmath648 by functoriality . this implies the monotonicity morphism @xmath649 is an isomorphism in grading @xmath650 since all three chain modules are @xmath0 in that grading . note that shrinking @xmath148 does not affect monotonicity morphisms except as generators of the respective chain modules appear or disappear . since when we shrink @xmath651 no closed reeb orbits of conley - zehnder index @xmath652 appear or disappear in chain modules @xmath653 and @xmath654 , @xmath650 we conclude that @xmath655 is an isomorphism too in this grading . because @xmath656 in all other gradings @xmath415 , @xmath193 is the only grading in which the chain map @xmath267 is non - trivial and by reasoning as in ( ii ) , one checks that the induced chain map @xmath657 consists of isomorphisms in row @xmath193 of the double complex @xmath658 and vanishes in all other rows . this implies @xmath659 is an isomorphism in all gradings @xmath660 . since this holds for all @xmath404 and @xmath606 sufficiently large it passes to the double limit , implying the inclusion morphism @xmath661 is an isomorphism for @xmath660 . to prove theorem [ thm : psi - perturbed ] we use the following : [ lem : zk - induction ] assume a @xmath0-chain map 0 & & _ k & ^c_0 ^ 1p & r & ^ ( t-1 ) & r & ^c_1 ^ 1p & & ^ ( t-1 ) & r & ^c_n^1p & & ^(t-1 ) & r&&0 + & & & & & & & & & & + 0 & & _ k & ^c_0 ^ 2p & r & ^ ( t-1 ) & r & ^c_1 ^ 2p & & ^ ( t-1 ) & r & & 0 where each arrow is multiplication by the specified field element @xmath662 , or polynomial @xmath663 , @xmath664 or @xmath665 for @xmath666 and @xmath667 , @xmath668 . then , if @xmath669 is a unit of @xmath0 , all @xmath670 , @xmath671 are units of @xmath216 . moreover , the conclusion also holds if the direction of the vertical arrows is reversed but all other hypotheses are the same . this follows by induction on @xmath672 using the commutativity of @xmath216 and the fact that @xmath516 , the maximal ideal of the local ring @xmath216 , contains all non - units and annihilates @xmath673 . fix @xmath404 . let @xmath674 be greater than both @xmath30 and @xmath395 and also sufficiently large that @xmath675 contains the support of @xmath17 . by theorem [ thm : comp - equiv ] the monotonicity morphism @xmath676 is an isomorphism . therefore by functoriality ( theorem [ thm : g - functor ] ) @xmath677 is an isomorphism and hence , using the fact these vector spaces are all equal to @xmath0 , @xmath678 is an isomorphism too . assume dominating sequences of smoothed , perturbed hamiltonians @xmath679 , @xmath680 and @xmath681 for @xmath7 , @xmath682 and @xmath675 respectively as in part ( iii ) of the previous proof where now the role of @xmath38 is played by @xmath683 for which @xmath684 $ ] . all contact forms @xmath603 , @xmath685 and @xmath602 therefore have a closed reeb orbit at @xmath309 with conley - zehnder index @xmath160 ; the forms @xmath602 have only this orbit , while the forms @xmath603 , @xmath685 have others as well . more precisely , from the discussion of morse - bott computations in the previous proof ( c.f . bourgeois @xcite ) the chain complexes @xmath686 and @xmath687 are of the form @xmath688 \to^{d_{1 } } c[-n-2n ] \to \cdots \to^{d_{\ell } } c[-n-2n\ell ] \to 0 $ ] for @xmath95 respectively given by @xmath528 $ ] or @xmath689 $ ] . moreover , by @xmath297-invariance , @xmath303 , @xmath690 is of the same form . by comparison with the known non - equivariant contact homology of the respective domains one deduces that for sufficiently large @xmath606 , all maps @xmath691 are multiplication by @xmath692 for @xmath517 and @xmath693 . by shifting the indexing of the sequence @xmath694 if necessary we may assume that @xmath695 , i.e. , @xmath696 . there is thus a corresponding induced @xmath0-equivariant chain map @xmath697 which will be as shown in lemma [ lem : zk - induction ] , but with all vertical arrows possibly directed upwards . since @xmath698 is an isomorphism and the vector spaces @xmath699 and @xmath700 are @xmath0 for all sufficiently large @xmath606 the chain map @xmath267 in degree @xmath160 must be an isomorphism for all sufficiently large @xmath606 and the result follows by applying lemma [ lem : zk - induction ] and passing to the double limit . by using / extending two of the constructions in eliashberg - kim - polterovich @xcite - namely , the map @xmath701 defined for @xmath702 ( see page 1649 of @xcite ) , @xmath703 and the squeezing given in theorem 1.19 of @xcite by a positive contractible loop of contactomorphisms of the ideal contact boundary @xmath704 - two observations follow . first , that non - squeezing past @xmath705 is equivalent to a squeezing room " requirement for squeezing past @xmath706 which is in some cases stronger than what is proved in @xcite are to be compared with statements in @xcite for @xmath707 ; we have used the letter @xmath708 instead of @xmath55 to avoid confusion with notation of preceding sections . considering the effect of the map @xmath709 , @xmath708 should be thought of as @xmath710 in this sentence . ] . second , even this stronger requirement may not be tight . to see the first observation , note that @xmath709 maps @xmath711 into @xmath712 for all @xmath713 and so in particular maps all of @xmath5 into @xmath714 , taking @xmath711 with @xmath715 to @xmath716 with @xmath717 . thus , theorem [ thm : main ] , resp . [ thm : main - strong ] is equivalent to in this sense , rigidity at large scale which completely precludes squeezing can be thought of as an infinite squeezing room requirement , and is equivalent to a form of rigidity at small - scale which requires squeezing room determined by the reciprocal integers . with this viewpoint , instead of a single cut - off between flexibility and rigidity of @xmath7 as @xmath30 grows , one sees rather a squeezing room requirement which jumps at each reciprocal integer , culminating in an infinite requirement when @xmath720 is passed . theorem 1.5 of @xcite established that there is no squeezing of @xmath719 into itself within @xmath721 but this bound depends on the particular @xmath722 . if equation holds and @xmath722 is not too close to @xmath723 - more precisely , if @xmath724 then we will have @xmath725 , i.e. , @xmath726 so the squeezing room requirement of theorem [ thm : main2 ] will ( for such @xmath722 ) be strictly stronger than that of theorem 1.5 of @xcite . we now remark that even this stronger squeezing room requirement is not necessarily tight . if one applies theorem 1.19 of @xcite to @xmath719 , @xmath727 as done in remark 1.23 of @xcite to @xmath728 , one obtains the more general : however , implies @xmath730 so there remains a gap between the required squeezing room and that of known squeezings . to find squeezings with smaller support it seems methods beyond the construction of @xcite would be needed , to produce either wilder " compactly supported contact isotopies which do not deform the original fiberwise star - shaped domain through fiberwise star - shaped domains , or compactly supported contactomorphisms not isotopic to the identity . besides closing the gap mentioned above another direction for future work is application of the framework of the present paper to prequantizations @xmath21 of other liouville manifolds @xmath121 . when @xmath121 is a sufficiently stabilized liouville manifold , squeezing at small scale has already been established by eliashberg - kim - polterovich ( c.f . theorems 1.16 , 1.19 in @xcite of which theorem 1.1 in our paper re - states the special case @xmath731 , @xmath732 ) . we conjecture that coarse non - squeezing at large scale holds as well in these settings . i would like to express my deep gratitude to leonid polterovich for his mathematical guidance during several semesters of my studies ( in computer science ) at the university of chicago , and for introducing me to the contact non - squeezing problem . none of the work in this paper would have started without the many discussions with him and this work became something of a second phd for me , bringing me back into mathematics . in particular the idea to tackle this problem by developing a @xmath0-equivariant form of contact homology is due to him . i owe a second debt of gratitude to isidora milin whose phd thesis on orderability of lens spaces was an invaluable resource to me in developing the @xmath0-equivariant theory in this paper and also revealed enlightening similarities between the seemingly unrelated non - squeezing phenomena considered in her setting and ours . i also thank frdric bourgeois for very helpful technical discusssions . finally i thank yasha eliashberg . his constant support and encouragement kept this project going despite countless delays . a portion of the writing of this work was completed while visiting tel aviv university in may 2014 and the universit de lyon in june - july 2014 . i thank both leonid polterovich and jean - yves welschinger for their hospitality and for stimulating discussions . those visits were partially supported by tau and the cnrs respectively . bourgeois - toronto f. bourgeois , _ a morse - bott approach to contact homology_. in symplectic and contact topology : interactions and perspectives " , eds . y. elashberg , b. khesin , f. lalonde , fields inst . commun . * 35 * , amer . soc.,2003 , pp . 55 - 77 . egh y. eliashberg , a. givental , and h. hofer , _ introduction to symplectic field theory_. gafa 2000 . in visions in mathematics " , eds . n. alon , j. bourgain , a. connes , m. gromov , v. milman , birkhuser , 2000 , pp . 560 - 673 . hwz h. hofer , k. wysocki , e. zehnder , _ properties of pseudoholomorphic curves in symplec- tizations . iii fredholm theory_. in topics in nonlinear analysis " , progr . nonlinear differential equations appl . * 35 , * birkhuser , 1999 , pp . 381 - 475 .

we define a @xmath0-equivariant version of the cylindrical contact homology used by eliashberg - kim - polterovich ( 2006 ) to prove contact non - squeezing for prequantized integer - capacity balls @xmath1 , @xmath2 and we use it to extend their result to all @xmath3 . specifically we prove if @xmath3 there is no @xmath4 , the group of compactly supported contactomorphisms of @xmath5 which squeezes @xmath6 into itself , i.e. maps the closure of @xmath7 into @xmath7 . a sheaf theoretic proof of non - existence of corresponding @xmath8 , the identity component of @xmath9 , is due to chiu ( 2014 ) ; it is not known if this is strictly weaker . our construction has the advantage of retaining the contact homological viewpoint of eliashberg - kim - polterovich and its potential for application in prequantizations of other liouville manifolds . it makes use of the @xmath0-action generated by a vertical @xmath10-shift but can also be related , for prequantized balls , to the @xmath0-equivariant contact homology developed by milin ( 2008 ) in her proof of orderability of lens spaces .

understanding interfacial phenomena at boundaries @xcite such as free surfaces of solids , or walls confining fluids , etc . , is a topic of very high interest in current condensed matter research , driven by important applications , e.g. in micro- or nanofluidics , design of smart nanomaterials @xcite , etc . but it is also of fundamental interest as a problem of statistical mechanics . it is only the latter aspect that will be in our focus here , and hence we shall not deal here with wetting properties @xcite of specific materials , but rather consider wetting in the framework of a simple lattice model with nearest neighbor interactions and short - range forces due to the walls . in this spirit , a large amount of work has been devoted to the study of simple ising models with boundary fields ; see e.g. @xcite for a review , and @xcite for some pertinent original work . spin reversal symmetry of this generic model implies that phase coexistence between the ordered phases with positive and negative spontaneous magnetization occurs at zero bulk field ( h=0 ) in the thermodynamic limit in the bulk ; the same holds still true when one considers a thin film of a finite thickness @xmath2 with surfaces at which boundary fields @xmath9 , @xmath10 act , provided one considers the antisymmetric situation @xmath11 , avoiding `` capillary condensation '' @xcite , i.e. a shift of bulk coexistence to nonzero value of the bulk field due to boundary effects . the latter phenomena is interesting in its own right , e.g. @xcite , and can also be studied in ising models choosing boundary fields @xmath12 , e.g. @xcite , but this problem will not be followed up here . in principle , the study of a wetting transition , where the boundary induces the formation of a macroscopically thick film of the phase that it prefers , coexisting with the other phase separated by an interface at a macroscopic distance from the boundary , requires to consider the limit @xmath13 . for any finite @xmath2 in the antisymmetric thin film geometry , or thin strip geometry , when one considers @xmath3 dimensions rather than @xmath14 , one does not find a wetting transition , but rather the " interface localization - delocalization transition @xcite : for temperatures above @xmath15 but below the transition temperature @xmath16 of the bulk ising model , the interface between the coexisting phases is delocalized " , freely fluctuating in the center of the film ( or strip , respectively ) . here , by center we mean the plane ( or line ) ( @xmath17 , if we label the layers ( rows ) of the ising lattice parallel to the boundaries from @xmath18 to @xmath19 . however , for @xmath20 the interface is tightly bound , and hence `` localized '' , near one of the two boundaries . one predicts , however , that for large @xmath2 the transition temperature @xmath21 converges rapidly to the wetting transition temperature @xmath22 of the semi - infinite system @xcite , and this prediction is consistent with the numerical simulations @xcite . one aspect of wetting in the ising model that found great attention is the behavior near bulk criticality . one can show that for a second - order wetting transition the inverse function @xmath8 of @xmath22 behaves as @xcite @xmath23 where @xmath24 is the critical exponent that controls the scaling behavior with the surface field @xmath9 near bulk criticality @xcite . while @xmath24 in @xmath14 is only known approximately from numerical work , in @xmath3 abraham s exact solution for the ising model @xcite implies @xmath25 . note that on the square lattice with exchange constant @xmath26 between nearest neighbors the wetting transition occurs for @xmath27 given as the solution of @xmath28\nonumber\\ & & = \sinh ( 2 j / k_bt ) \quad , \end{aligned}\ ] ] and the bulk critical point occurs at @xmath29 @xcite . an intriguing question is , of course , to study how the wetting behavior gets modified if the bulk phase transition changes its character . e.g. , a simple extension of the ising model is the blume - capel model @xcite , where each lattice site ( i ) can be in one of three states @xmath30 , @xmath31 , @xmath32 and the hamiltonian becomes @xmath33 where the `` crystal field '' @xmath34 controls the density of `` vacancies '' , i.e. sites with @xmath31 . when @xmath35 vacancies are excluded , and the model reduces to the standard two state ising model , but when @xmath36 is of order unity , one has a nontrivial bulk phase diagram with a tricritical point @xcite at @xmath37 . for the square lattice , this tricritical point occurs at @xmath38 and @xmath39 ( 5 ) . for @xmath40 the transition at @xmath41 is first order : two phases with opposite nonzero magnetization and the disordered phase ( with zero magnetization ) coexist . from the universality principle , we expect that eq . ( [ eq1 ] ) with @xmath25 would also hold in the second - order region of the blume - capel model for @xmath42 , but to our knowledge this assumption has not yet been tested . providing evidence for this assumption is one of the goals of the present paper . right at the tricritical point , however , the behavior should be different : instead of eq . ( [ eq1 ] ) we now expect @xmath43 where the exponent @xmath44 in @xmath14 is expected to be @xmath45 @xcite , the prediction due to landau mean - field theory , which is expected to hold in @xmath14 , apart from logarithmic corrections @xcite . while in @xmath3 the tricritical exponents in the bulk are not those of mean - field theory @xcite , for the ising model universality class to which the tricritial point of the blume - capel model belongs @xcite , they are exactly known from conformal invariance @xcite . e.g. in standard notation of critical exponents @xcite ) @xmath46 , @xmath47 , @xmath48 , @xmath49 , etc . to our knowledge , @xmath44 for this problem is not yet known , however . when one considers the wetting transition for the part of the transition line @xmath50 that is first order @xmath51 , the wetting transition @xmath8 ends at this line at a nonzero value of the surface field , @xmath52 we expect that another power law will exist when @xmath34 approaches @xmath53 from below , @xmath54 but again we are not aware of predictions relating to this critical exponent @xmath55 . the evaluation of @xmath55 from numerical work is a very challenging task , because it requires to locate both the tricritical point and the wetting transitions with very high accuracy . finally , a very interesting aspect relates to the character of the interface between the coexisting phases with positive and negative magnetization , when one approaches the bulk phase transition : then interfacial adsorption of the third phase ( the disordered phase ) at the interface can occur @xcite . if we denote the coexisting bulk phases as @xmath56 and @xmath57 , approaching the bulk transition line in the second - order region one expects an interfacial wetting transition @xmath58 , where @xmath59 stands symbolically for an intruding layer of the disordered phase having predominantly states where @xmath31 dominates at the interface . to quantify this effect one considers the net adsorption defined by @xmath60 \quad , \ ] ] where @xmath61 is the total number of lattice sites , @xmath62 is the number of lattice sites per row parallel to the walls , @xmath63 is the kronecker symbol , and @xmath64 a statistical average in the presence of an interface ( @xmath65 , @xmath66 ) , while @xmath67 is the corresponding average for an equivalent system but without interfaces . since @xmath68 , one can conclude also that @xmath69 where @xmath70 is the interfacial tension for the considered interface . since for @xmath42 the bulk phase transition of the blume - capel model falls in the ising universality class , we know that @xmath71 with @xmath72 for the blume - capel model . writing @xmath73 , we find @xmath74 thus if we define a divergence of interfacial adsorption via @xmath75 we find @xmath76 for @xmath42 , i.e. in the ising - like regime there should be no critical divergence of the net adsorption . however , right at the tricritical point , the behavior should be different : from @xmath77 as quoted above we conclude @xmath78 with @xmath79 for @xmath37 . this result is consistent with monte carlo results of selke et al . @xcite . for @xmath80 the transition of the blume - capel model is believed to be first order , and the emerging interfaces @xmath81 and @xmath82 remain sharp as @xmath83 . reducing the problem to interfaces in the solid - on - solid model , one can argue that eq . ( [ eq9 ] ) also holds in the first - order region , but @xmath84 @xcite . these considersations have been checked by early monte carlo work @xcite for interfaces between coexisting bulk phases , but no study in the context of wetting at external boundaries has as yet been performed , to our knowledge . in the present work , we wish to contribute filling this gap and study wetting behavior for the two - dimensional blume - capel model in the case of thin films for which surface fields at the boundaries act , as described in sec . 2 . however , in order to do so , we have found it necessary to reconsider the simulation methodology for the study of critical wetting . in fact , by applying monte carlo simulations , we shall necessarily study finite systems , namely strips of width @xmath2 ( in @xmath85-direction ) and length @xmath62 ( in @xmath86-direction , where a periodic boundary condition acts ) . as a consequence , finite size effects matter , and hence in the next section we also shall give the background on the proper finite size scaling analysis of such simulation `` data '' , for the case of critical wetting in @xmath3 dimensions , taking into account that one deals there with an anisotropic critical phenomenon of a special character : if we consider the magnetization @xmath87 of the strip as the `` order parameter '' of the transition , its critical exponent @xmath88 . this fact does not seem to have found much attention in the previous monte carlo studies of critical wetting in @xmath3 @xcite . this revised methodology for the study of critical wetting , which is outlined in sec . 2 , is another important result of our work . sec . 3 then presents our monte carlo results , which then are analyzed according to these concepts , and the wetting transitions are discussed for a variety of choices of @xmath34 , both cases @xmath42 , @xmath37 and @xmath89 will be considered . 4 then discusses our result on interfacial adsorption , and sec . 5 summarizes our conclusions . we consider the hamiltonian of the @xmath90-state blume - capel model @xcite where each lattice site @xmath91 carries a spin @xmath92 that can take on the values @xmath93 at a square lattice in a @xmath0 geometry , where periodic boundary conditions act in the @xmath86-direction ( where the lattice is @xmath62 rows long ) , while free boundary conditions are used in the @xmath85-direction , where boundary fields @xmath9 , @xmath10 act on the first and the last row . thus the hamiltonian is @xmath94 here @xmath26 is the exchange constant between spins at nearest neighbor sites , which we take homogeneous throughout the system . thus , we disregard the possibility to take the exchange @xmath95 in the boundary rows different from the exchange @xmath26 in the interior of the system , that is frequently considered in studies of wetting phenomena in the ising model @xcite . the bulk field @xmath96 acts on all lattice sites , while the `` surface fields '' @xcite only act on the spins in the first ( 1 ) and last ( l ) row , where the free boundary conditions apply . we also specialize on the particular antisymmetric situation @xmath97 and consider the thermodynamic limit @xmath98 , @xmath99 ) before we consider the limit @xmath100 . then , the system undergoes two phase transitions : at the temperature @xmath50 the phase transition occurs from the disordered `` paramagnetic '' phase to the ordered `` ferromagnetic '' phase , where we have used the terminology of the ising model . taking the lattice spacing as our unit of length , the total number of spins is just the product of the linear dimensions of the system , @xmath101 . in fact , when we define the magnetization @xmath87 per lattice site as @xmath102 its thermal expectation value @xmath103 for temperatures @xmath104 will be nonzero and positive in the considered limit . however , a second phase transition occurs at a lower temperature @xmath22 for small enough absolute values @xmath105 of the surface field : note that the surface field @xmath9 is oppositely oriented to the positive bulk field @xmath96 , but @xmath100 while @xmath9 , stays finite . thus , near @xmath106 the surface field stabilizes a macroscopically thick layer of negative magnetization near the lower boundary , where @xmath107 acts , separated by an interface from the bulk , where the magnetization is positive . at @xmath108 , a transition occurs where this interface gets localized near the lower boundary : in the extreme case , the domain with negative magnetization disappears completely . in the ising model , which results from eq . ( [ eq10 ] ) as the limiting case @xmath109 , this wetting transition is second order throughout the regime @xmath110 . for the semi - infinite system described above , the wetting transition is a singularity of the surface excess free energy @xmath112 , defined from standard decomposition of the total free energy into the bulk term and boundary terms , for @xmath13 , @xmath99 @xmath113 the singular part of this boundary free energy , also called `` wall tension '' or `` wall excess free energy '' @xcite , near @xmath22 is expected to satisfy a scaling behavior @xcite where @xmath114 , @xmath115 @xmath116 being a scaling function that we do not specify here . this is analogous to standard scaling in the bulk near @xmath16 @xcite , where @xmath117 , @xmath118 in eq . ( [ eq14 ] ) , @xmath119 is the specific heat exponent , and the `` gap exponent '' @xmath120 is related to the standard exponents @xmath121 of the spontaneous magnetization @xmath122 and @xmath123 of the susceptibility @xmath124 @xcite . thus @xmath125 , @xmath126 are analogous exponents characterizing the singular behavior of the critical wetting transition in @xmath3 dimensions . as always when one considers a critical phenomenon , a diverging correlation length exists . since wetting can be viewed as an `` interface unbinding '' transition @xcite when approached from below , it is natural to study the correlation function @xmath127 describing the correlation of fluctuations of the height of the contour @xmath128 separating the domain with negative magnetization near the lower boundary from the region with positive magnetization in the bulk . defining @xmath129 as @xmath130 , this yields @xmath131 noting translational invariance in @xmath86-direction . by @xmath132 we denote an average in the canonical ensemble at temperature @xmath133 ; in the following the subscript @xmath133 will be omitted . near critical wetting in @xmath3 @xmath134 takes the scaling form @xcite @xmath135 writing analogously to the decay of bulk correlations near criticality , @xmath136 , with @xmath137 , @xmath138 being appropriate exponents describing the variation of these correlations at large distances right at bulk criticality or critical wetting , where @xmath139 ( or @xmath140 is infinite . for @xmath141 , a scaling relation analogous to eq . ( [ eq13 ] ) holds , @xmath142 again this ansatz is inspired by the corresponding bulk behavior , @xmath143 and again @xmath144 , @xmath145 are scaling functions that we do not specify here . we disregard here the problem that in general a nonlocal theory @xcite is required @xcite , since it does not change the following conclusions . capillary wave theory implies that @xmath146 , so there is no second independent critical exponent as in the bulk . for a more detailed reasoning to explain why for the case of critical wetting with short - range forces the exponent @xmath138 is zero in all dimensions , and hence there is a single independent critical exponent , we refer to the literature @xcite . remember that using hyperscaling relations @xmath147 and the scaling law @xmath148 one can express all static critical exponents just in terms of only two independent exponents @xmath149 and @xmath137 in the bulk @xcite . for critical wetting , in @xmath3 , there is only one independent critical exponent namely , @xcite @xmath150 and the hyperscaling relation for interfacial phenomena in @xmath3 dimensions @xmath151 @xcite then implies @xmath152 . from eq . ( [ eq16 ] ) we then conclude that on a length scale @xmath86 the interface exhibits ( at @xmath153 ) a mean - square displacement also proportional to @xmath86 , while for @xmath154 one concludes that on the length scale @xmath141 the mean - square displacement in @xmath85-direction also is proportional to @xmath141 , while for @xmath155 the correlation @xmath156 . this leads to the conclusion that the correlation length @xmath157 describing the fluctuations of the interface in @xmath85 direction scales as @xmath158 noting from eq . ( [ eq13 ] ) that surface excess magnetization @xmath159 and surface excess susceptibility @xmath160 follow from derivatives with respect to the field @xmath161 @xmath162 we conclude that these quantities have the singularities @xmath163 and @xmath164 using then the consideration that this excess susceptibility is just caused by displacements of the interface between the domain of negative magnetization at the boundary and the bulk , one can use eq . ( [ eq16 ] ) to deduce a further scaling relation between @xmath165 @xmath166 and @xmath138 , namely @xcite @xmath167 = 3 , \quad ( d=2 ) \quad.\ ] ] also the excess magnetization @xmath159 obviously simply is related to @xmath168 , and we have , in @xmath3 , @xmath169 thus , near to critical wetting in @xmath3 the typical mean distance of the interface from the wall and the typical excursions of this interface from its mean are of the same order , @xmath170 . all above relations did refer to the case that we were considering the limits @xmath99 and @xmath13 first , and then consider the limits @xmath171 , @xmath172 near the wetting transition . while such an approach is natural in the context of analytic theories @xcite , it it not sufficient to understand simulations , where we wish to take these limits in reverse order . when @xmath2 is kept large but finite , it is clear that for temperatures above the wetting transition and @xmath173 there is no physical distinction between the domain with negative magnetization in the lower half of the system and the domain with positive magnetization in the upper half : previously , for the semi - infinite case , the positive magnetization in the bulk was singled out by taking the limit @xmath100 first , and then the positive boundary field @xmath10 for @xmath13 was irrelevant , while for @xmath2 finite it is critical to maintain the symmetry of the situation with respect to the sign of the magnetization for @xmath173 . as a consequence , we conclude that for the present situation there is no bulk magnetization to consider , and it rather is the total magnetization of the system or strictly speaking , its absolute value , which undergoes a transition from zero to a nonzero value when we cross the wetting transition temperature , and extrapolate simulation results towards @xmath13 . as a consequence of this consideration , we propose a scaling assumption for the distribution function @xmath174 of the total magnetization in this finite geometry as follows @xcite @xmath175 note that this expression generalizes the standard expression for finite - size scaling in isotropic systems that have linear dimension @xmath2 in all spatial directions @xcite , as appropriate for phase transitions in the bulk @xmath176 to systems with anisotropic linear dimensions @xmath177 and anisotropic correlation length exponents @xmath166 , @xmath178 . while for isotropic critical phenomena an anisotropic system shape would lead to a dependence simply on the `` aspect ratio '' @xmath179 of the system , the fact that @xmath62 scales with @xmath141 and @xmath2 scales with @xmath157 can be used to demonstrate @xcite that the @xmath2-dependence enters via `` the generalized aspect ratio '' @xmath180 in the scaling function , eq . ( [ eq27 ] ) . with @xmath181 , @xmath182 we simply have an argument @xmath183 , which needs to be kept constant when the variation with @xmath62 is studied . such generalizations of finite - size scaling to anisotropic criticality have been discussed earlier for the kasteleyn transition @xcite , which also exhibits @xmath184 and hence the same generalized aspect ratio @xmath185 applies . the prefactor @xmath186 in eq . ( [ eq27 ] ) , and likewise the prefactor @xmath187 in eq . ( [ eq28 ] ) , ensures that the probability @xmath188 can be properly normalized , @xmath189 taking now suitable moments of @xmath188 , analogous to the procedure at isotropic phase transitions in the bulk @xcite , we obtain @xmath190 @xmath191 in particular , from eqs . ( [ eq30 ] ) , ( [ eq31 ] ) we derive the standard behavior for the `` susceptibility '' . we denote it by @xmath192 since @xmath193 rather than @xmath194 is subtracted , see ref . @xcite for a discussion . so , one has @xmath195 and hence @xmath196 with the scaling function @xmath197 , omitting all arguments for simplicity . now the key task is to identify the `` order parameter exponent '' @xmath121 in the context of this description of a critical wetting transition . for this purpose , we note that the singular behavior of @xmath192 can only be due to the singular behavior of @xmath198 , eq . ( [ eq24 ] ) , where we used the results @xmath152 , @xmath199 . a key point to note is that @xmath192 was normalitzed per spin , relating to the total volume @xmath200 , while @xmath160 in eq . ( [ eq24 ] ) is taken relative to boundary sites , so is only normalized by @xmath62 . hence we conclude that at @xmath201 , we have , using @xmath202 in eq . ( [ eq33 ] ) to eliminate @xmath2 @xmath203 irrespective of how the finite constant @xmath204 is chosen . writing then @xmath205 and using a finite - size scaling relation for @xmath160 @xmath206 and hence we find @xmath207 comparison of eqs . ( [ eq34 ] ) and ( [ eq37 ] ) yields a central result of this section , namely @xmath208 and eq . ( [ eq37 ] ) also can be interpreted in terms of the standard finite - size scaling result @xmath209 second order transitions with an exponent @xmath210 are rather unusual ; for another recent example see jaubert et al . @xcite . in the present case , we can also understand it from the fact that @xmath211 must tend to @xmath212 in the thermodynamic limit in the partially wet phase , where the interface is bound to one of the walls up to the transition point , while in the wet phase for the considered limit @xmath213 . it is interesting to note that these critical exponents also satisfy the usual scaling relation with the gap exponent @xmath126 introduced in eq . ( [ eq13 ] ) , @xmath214 as well as the hyperscaling relation for anisotropic bulk critical phenomena in @xmath111 dimensions , @xmath215 one should not confuse the exponent @xmath121 of the total magnetization , resulting in the limit @xmath13 , @xmath99 , @xmath216 , with the exponent @xmath217 of the surface excess magnetization @xmath159 defined in eq . ( [ eq23 ] ) : while @xmath218 as @xmath172 , @xmath159 diverges as @xmath172 . actually , we have the standard scaling relations @xcite @xmath219 with @xmath220 , as is easily checked by noting that @xmath221 . a short manipulation of eq . ( [ eq33 ] ) shows that it can be rewritten as @xmath222 where @xmath223 is another scaling function . for a fixed generalized aspect ratio @xmath224 this function has a maximum at same value @xmath225 of the argument @xmath226 . this implies that the height of the maximum scales as @xmath227 in the present case ) ; see also eq . ( [ eq37 ] ) ; and its position @xmath228 scales as @xmath229 incidentally , this extrapolation has already tentatively been used in an early work @xcite without detailed justification . the interpretation of the wetting transition in analogy to bulk critical phenomena , considering a particular way of taking the thermodynamic limit ( cf . [ fig1 ] ) , is reminiscent of studies of filling transitions in double wedge @xcite or bipyramid @xcite geometries ; however , there the critical exponents are different . a useful consequence of eq . ( [ eq38 ] ) is that both @xmath211 and @xmath230 right at the critical wetting transition become completely independent of linear dimensions , but depend still on the constant @xmath231 , the generalized aspect ratio . plotting @xmath211 vs. temperature for different choices of @xmath62 at constant @xmath204 , critical wetting should show up via an unique intersection point . the fact that for wetting transitions in @xmath3 for the choice @xmath232 a unique intersection point of @xmath233 versus @xmath133 curves occurs has been only noted previously @xcite in a study where long range boundary fields were applied , where @xmath234 holds @xcite , and hence the critical behavior is rather anomalous . [ fig1 ] shows a schematic sketch where the typical excursion @xmath157 of the interface is still smaller than @xmath2 , by a factor of about 2.5 , and similarly @xmath141 is smaller than @xmath62 . the absolute value @xmath211 of the magnetization then still is of the same order as the bulk magnetization @xmath235 that one encounters inside of the domains . the distribution of @xmath87 then is distinctly bimodal , and @xmath236 , since the interface is bound to the upper wall with the same probability as the lower wall . however , when @xmath133 approaches @xmath201 , the attractive force between the wall and the interface is compensated by the entropic repulsion , the excursions of the interface extend over the whole width @xmath2 , and then show a correlation over a lateral length @xmath237 . this interface configuration at @xmath22 then is selfsimilar , irrespective how large @xmath2 is chosen , as expected for critical phenomena . thus , the result @xmath88 must not be confused with a first - order wetting transition , where one would have a discontinuous change from a bound to an unbound state of the interface , rather than this continuous unbinding . of course , it is of interest to ask how the behavior changes when one does not take the thermodynamic limit in this particular way where the generalized aspect ration @xmath185 is fixed , but rather choosing one of the linear dimensions large but finite , and varying only the magnitude of the other linear dimension . thus , if @xmath2 is kept fixed and one considers the limit @xmath99 , the system becomes quasi - one- dimensional . while mean - field theory would predict a sharp interface localization - delocalization transition in this limit @xcite , one expects that thermal fluctuations cause a rounding of this transition since the system is equivalent to a one - dimensional ising system @xcite , and this expectation has in fact been verified numerically @xcite . if instead we keep @xmath62 finite and increase @xmath2 , the growth of @xmath141 would be limited by the magnitude of @xmath62 , and hence also @xmath157 as well as the distance of the interface from one of the walls in the bound state would be limited by a value proportional to @xmath238 , irrespective of @xmath2 , while in the wet state its equilibrium position is at @xmath239 . as a consequence , data taken in such a way would appear as if the wetting transition were weakly of first order , which it is not . thus , we conclude that only the finite - size scaling analysis at constant generalized aspect ratio @xmath240 is the most useful choice . in most of the numerical work , we have chosen a particular value of this generalized aspect ratio , namely @xmath241 of course , the value of the constant @xmath204 in principle is arbitrary , and the results on the location of @xmath22 and the critical exponents should not depend on this choice . our motivation for this choice was simple that it yields solutions for @xmath2 and @xmath62 which are both integer , as it must be on the lattice , and rather small : @xmath242 ; and @xmath243 . however , the effect of considering other choices of @xmath204 will be discussed below . monte carlo simulations were then performed using the standard metropolis algorithm , see e.g. @xcite for a review . typical runs are performed over a length of 10@xmath244 monte carlo steps per lattice site ( mcs ) , disregarding the first @xmath245 mcs to allow the system to reach equilibrium . note that for systems far below bulk criticality exposed to boundary fields cluster algorithms do not present any advantage @xcite . in addition , part of the bulk phase diagram of the blume - capel model has been calculated , by applying again standard monte carlo methods @xcite but using finite - size scaling for @xmath246 lattices with periodic boundary conditions in both @xmath86 and @xmath85 directions , see fig . [ fig2 ] . for earlier studies see e.g. @xcite . motivated by the analysis of sec . ii , we have analyzed the first two moments @xmath233 , @xmath247 and the cumulant @xcite @xmath248 \quad , \ ] ] of the magnetization distribution and plot them versus temperature @xmath133 for a few typical values of @xmath36 , as shown in figs . [ fig3]- [ fig5 ] ) . as expected , rather well - defined intersection points in all three quantities @xmath233 , @xmath249 and @xmath250 can be found at the same estimate @xmath251 within reasonably small errors . as always @xcite , the statistical accuracy is better for @xmath233 and @xmath249 than for the cumulant . for standard phase transitions in the bulk @xcite , @xmath252 and @xmath249 would not exhibit `` universal '' , size - independent , intersection points , of course . so , the behavior seen in figs . [ fig3]- [ fig5 ] is clear evidence for the scaling description of sec . ii , and also the direct observation of configuration snapshots , see e.g. fig . [ fig6 ] , is compatible with the qualitative picture that was developed ( fig . [ fig1 ] ) . note that in the wet phase the average position of the interface is in the middle of the @xmath0 strip , in between the rows @xmath253 and @xmath254 , for @xmath2 even as chosen here ; but due to capillary waves on the scale @xmath62 in @xmath255direction the interface makes excursions of order @xmath238 in @xmath256direction , which are of the order of @xmath239 , for our choice of geometry ( eq . ( [ eq43 ] ) ) . on the other hand , in the case shown for @xmath257 , i.e. the ising limit of the blume - capel model ( fig . [ fig3 ] ) , the location of the wetting transition is exactly known @xcite , eq . ( [ eq2 ] ) , and highlighted by a vertical straight line . evidently , the intersection of the curves in fig . [ fig3 ] occurs at a temperature compatible with this prediction . we do expect that the critical wetting transition of the blume - capel model falls in the same universality class as for the ising model . thus , for the same choice of the generalized aspect ratio @xmath204 , eq . ( [ eq43 ] ) , we expect the same ordinate values of the intersection points of @xmath233 , @xmath249 and @xmath250 , respectively . within our accuracy , the data are compatible with this expectation . [ fig7 ] analysis the scaling of the peak heights and peak locations of the susceptibility @xmath258 , cf . ( [ eq37 ] ) and eq . ( [ eq33c ] ) , respectively . also , figs . [ fig8 ] and [ fig9 ] present some examples for the `` data collapse '' on master curves when @xmath259 , @xmath247 and @xmath260 are plotted as a function of @xmath261 as a test of eq . ( [ eq30 ] ) , @xmath262 , @xmath263 ) where we used the asymptotic power law for @xmath141 , @xmath264 . one sees that a fair data collapse does in fact occur , although deviations due to both statistical errors and systematic effects are present , when @xmath62 and/or @xmath141 are not large enough or when data too far away from @xmath22 are included . of course , figs . [ fig7]-[fig9 ] are just examples only , but representative for the general pattern of behavior . in any case , we do conclude that using such finite - size scaling analyses as presented here one can locate wetting transitions for two dimensional lattice models , such as the blume - capel model on the square lattice , with reasonable accuracy . on view of the well - defined intersection points and the data collapse in all three quantities @xmath233 , @xmath249 and @xmath250 so far studied , figures [ fig3]-[fig5 ] and [ fig8]-[fig9 ] , respectively , it is worth to analyze the scaling behavior of the distribution function @xmath174 of the total magnetization given by equation ( [ eq27 ] ) . in fact , since one has @xmath210 the prefactor and the third scaling argument are constants . also , just at @xmath251 ( i.e. @xmath265 ) the second scaling argument vanishes due to the divergence of the correlation length @xmath141 , so that the distribution function depends on the generalized aspect ratio @xmath266 , equation ( [ eq43 ] ) , of the sample only . figures [ fig10 ] ( a ) and ( b ) show plots of @xmath174 versus @xmath87 as obtained by keeping @xmath267 and for the cases @xmath268 and @xmath269 , respectively . it might be expected that the curves of the probability distribution would be independent of @xmath34 , except for normalization corrections due to the different density of vacancies . since we found that the shape of the probability distribution depends sensitively on @xmath133 , it could be that the qualitative differences between the curves shown in figures [ fig10 ] ( a ) and ( b ) would be due to the uncertainties in the location of @xmath251 for the case @xmath269 . on the other hand , by keeping the temperature , the crystal field , and the surface magnetic field just at the wetting transition point , we have measured the dependence of @xmath174 on the generalized aspect ratio , for the ising model , as shown in figure [ fig11 ] ( a ) . in that figure we choose @xmath270 while @xmath62 is varied . we recall the non - trivial shapes of the probability distributions that exhibit a single peak around @xmath271 for rather elongated samples ( i.e. @xmath272 ) , which monotonically crosses over to a bimodal distribution in the limit of more cubic samples ( i.e. @xmath273 ) . also , the generalized aspect ratio selected for our detailed monte carlo simulations ( @xmath267 ) , which can roughly be located within the crossover regime , exhibits not only the central peak , but it also shows the onset of growth of lateral shoulders close to @xmath274 . by using the distributions already shown in figure [ fig11 ] ( a ) we can also measure the dependence of the crossing points on the generalized aspect ratio of the sample ( figure [ fig11 ] ( b ) ) . from this plot it follows that the suitable range of @xmath204 in order to obtain a reliable intersection point for both @xmath233 and @xmath250 in monte carlo simulations is rather narrow , say @xmath275 , and slightly broader for the case of @xmath249 . in order to be able to locate intersections with reasonable accuracy , it is advisable that the cumulant intersection is neither close to zero nor close to its maximun value ( @xmath276 ) . it is also reassuring to see that for no shape the distribution at the wetting transition resembles the shape expected for a first - order wetting transition , which would be three delta functions at zero and positive and negative spontaneous magnetization in the thermodynamic limit , and a distribution with three peaks of approximately gaussian shape in a large but finite system . the location of the wetting transitions as a function of two parameters , @xmath277 and @xmath36 , by determining well - defined intersection points still requires extensive computations and hence is a challenging task , figs . [ fig12 ] and [ fig13 ] . since @xmath278 at @xmath279 decreases only rather slowly with increasing @xmath34 when one approaches @xmath53 , it obviously is very difficult to estimate the exponent @xmath44 ( eq . ( [ eq4 ] ) numerically . in mean - field theory , supposed to be exact for @xmath280 we know that @xmath281 and @xmath45 @xcite ; while in @xmath14 dimensions numerical estimates yielded @xcite @xmath282 , and we know that in @xmath3 @xmath281 again . so , we see that the exponent @xmath24 never deviates much from its mean - field value , if at all . if we speculate the same observation to be true for @xmath44 , we would expect that the data in fig . [ fig13 ] for @xmath283 vary as @xmath284 , which is not unreasonable . we also see that in the region where the transition in the bulk is first order , the wetting transition lines @xmath285 indeed end at nonzero values @xmath286 at the bulk transition line , and these values decrease as one approaches the tricritical point , qualitatively compatible with eq . ( [ eq6 ] ) . thus , we can sketch the global wetting behavior as shown schematically in fig . [ fig14 ] ; however , it would be premature to attempt to estimate the exponent @xmath287 : our data are clearly too limited for this purpose , and we have not tried to extend our study , since we feel the incomplete knowledge on the precise location of the bulk tricritical point @xcite would hamper such a study . as a final problem we consider the enrichment of vacancies at the interface between oppositely magnetized domains in the blume - capel model . this problem has been considered earlier by selke et al . @xcite for the case of free unbound interfaces , while here we consider the extension where the interfaces are confined between competing walls and may undergo a wetting ( interface unbinding ) transition . [ fig15 ] shows plots of the profiles of the vacancy density ( a ) and the magnetization ( b ) for a situation of incomplete wetting . one observes a clear enrichment of the vacancies in the interfacial region . the more the interface unbinds from the wall , the more the peak of the vacancy concentration moves away from the wall ( figs . [ fig15 ] a , b ) . note that for the low temperature shown , the magnetization in the bulk tends to @xmath288 , i.e. the vacancy concentration inside the bulk is very small . close to the wetting transition , the interface gets detached from the wall , and then the interface is located in the middle of the strip , at @xmath289 on average . however , since the interface strongly fluctuates around its average position a straightforward measurement of the vacancy concentration profile would yield an almost horizontal flat curve across the strip . so we have defined a local coarse - graining of the interface , taking segments of length @xmath290 in the @xmath86-direction , and determining the local center of mass @xmath291 of the interface in each segment , we compute the vacancy distribution in each segment separately , relative to the center in each segment , and superpose the distributions from the individual segments such that their centers coincide . in this way one obtains vacancy profiles with a clear peak in the center of the strip in the wet phase ( fig . [ fig15]c ) . finally , when one reaches the phase boundary of the bulk , the disordered phase takes over in the film , and the vacancy concentration becomes ( almost ) unity in the system , except near the boundaries where the surface fields still stabilize layers of up - spins and down - spins , respectively ( fig . [ fig15]d ) . note , of course , that the details of the curves in fig . [ fig15]c , d do depend on this coarse - graining length @xmath292 distinctly . this ambiguity that results depend on the coarse - graining is avoided when one simply computes the average fraction of vacancies in the system ( fig . [ fig16 ] ) . while for @xmath105 chosen such that the system is in the incompletely wet state up to the transition in the bulk one sees also a clear jump in the vacancy concentration from a small value to almost unity when this transition in the bulk occurs , a much more rounded behavior is found when wetting at the boundaries occurs . the difference between incompletely wet states and completely wet states is even more pronounced when one studies the magnetization fluctuation : for wet states , this fluctuation is very large for almost all values of @xmath34 until the transition in the bulk is reached ( fig . [ fig16]b ) . this behavior is corroborated by an examination of snapshot pictures ( fig . [ fig17 ] ) . finally , fig . [ fig18 ] shows an attempt to study the divergence of the interfacial adsorption , that we expect when we approach the bulk transition in the wet phase in the region where a first - order transition occurs . however , it is seen that the effective slope of these data changes slightly when the transition is approached , and also depends on @xmath293 . as a consequence , we must conclude that the asymptotic region where a universal exponent could be estimated has not been reached . in view of the strong and slow fluctuations of the interface it is not straightforward to cope with this problem . in this paper , we have presented a monte carlo study of the wetting behavior of the two - dimensional blume - capel model , mapping out the surface critical field @xmath294 where critical wetting occurs for a broad range of parameters @xmath133 and @xmath34 , both in the region where the transition of the blume - capel model in the bulk is of second order ( cf . [ fig12 ] ) and where is of first order ( cd . fig . [ fig13 ] ) . we suggest that @xmath295 behaves as @xmath296 for all @xmath297 and @xmath298 for @xmath299 , and propose a qualitative phase diagram in the space of the three variables surface field ( @xmath300 ) , temperature ( @xmath133 ) and `` crystal field '' ( @xmath34 ) , as sketched in figure [ fig14 ] . thus , we present evidence for the universality of the surface critical exponent @xmath301 along the critical line of the blume - capel model . however , the precise crossover behavior in the inmediate vecinity of the tricritical point @xmath302 could not yet been studied : first of all , the location of this point needs to be more accurately determined , and also more powerful numerical methods ( rather than the straightforward monte carlo simulation method just using the metropolis algorithm ) would be required . already for the task that we did achieve we encountered the necessity to reconsider the finite - size scaling approach to the study of wetting transitions , sec . we propose that by using a procedure where the linear dimensions @xmath303 of the @xmath0 strip are varied such that the generalized aspect ratio @xmath304 is kept constant , cf . ( [ eq43 ] ) , it is possible to analyze the data in full analogy to the study of a bulk phase transition , where one uses antisymmetric surface fields ( @xmath305 ) . then , simply the total magnetization @xmath87 in the system acts like as order parameter , but the appropiate critical exponent is @xmath210 . using the exactly known results for the ising model in the square lattice , which is the limit of the blume - capel model for @xmath306 , as a test case , our new finite - size scaling approach is nicely verified . note that @xmath307 and the critical exponents @xmath308 , @xmath309 , @xmath310 , @xmath311 , @xmath312 , and @xmath313 for critical wetting are known . then , we propose the corresponding exponents when one treats the transition not as a surface free energy singularity of the semi - infinite system in the limit where the bulk field @xmath96 tends to zero , but as a bulk singularity of the @xmath0 system for @xmath171 in the limit @xmath314 , @xmath315 constant . these expnents are @xmath316 , @xmath210 and @xmath317 . we have shown that these exponents satisfy all expected scaling relations and we have verified them from our simulations . use of this formulation of finite - size scaling for critical wetting in @xmath111 dimensions is a convenient and useful tool ; it would be interesting to study the extension of this method to @xmath318 dimensions where @xmath319 ( logarithmic growth ) and this extreme shape anisotropy that our method then requires makes the task obviously very much harder . since the blume - capel model differs from the ising model by the presence of vacancies as a third component , it is interesting to ask to what extent vacancies get enriched at the interfaces in this situation of interfaces confined by walls . we found that the situation is fully analogous to interfaces between coexisting phases in the bulk : in the second - order region the interfacial adsorption is finite , while in the first - order region the predicted divergence of interfacial adsorption is found . of course , there are many cases where by changing a parameter the order of a phase transition in the bulk changes from second to first order , and one can ask how wetting phenomena are affected . we hope that the present work will stimulate further theroretical and experimental studies along such lines . * acknowledgement * : one of as ( eva ) received support from the alexander von humboldt foundation and from the schwerpunkts fr rechnergesttzte forschungsmethoden in den naturwissenschaften ( srfn ) , germany . also , the support of the the conicet , anpcyt and unlp ( argentina ) is greatly acnowledged . 0 j.s . rowlinson and b. widom , _ molecular theory of capillarity _ ( oxford university press , oxford , 1982 ) p.g . de gennes , f. brochard - wyart , and d. quer : _ capillarity and wetting phenomena : drops , bubbles , pearls , waves _ ( springer , berlin , 2003 ) m. schn and s. klapp , _ nanoconfined fluids : soft matter between two and three dimensions _ ( j. wiley & sons , new york , 2006 ) i. brovchenko and a. oleinikova , _ interfacial and confined water _ ( elsevier , amsterdam , 2008 ) g. decker and j.b . schlenoff ( eds . ) _ multilayer thin films : sequential assembly of nanocomposite materials _ ( wiley - vch , weinheim , 2002 ) y. champion and h .- j . fecht ( eds . ) _ nano - architectured and nano - structured materials _ ( wiley - vch , weinheim , 2004 ) s. zhang ( ed . ) _ handbook of nanostructured thin films and coatings , vol . 1 - 3 _ ( crc press , boca rotan , 2010 ) t.m . squires and s.r . quake , rev . phys . * 77 * , 977 ( 2005 ) p.g . de gennes , rev . * 57 * , 827 ( 1985 ) d.e . sullivan and m.m . telo da gama , in c. croxton ( ed ) _ fluid interfacial phenomena _ ( wiley , new york , 1986 ) p.45 d.b . abraham , in c. domb and j.l . lebowitz ( eds . ) _ phase transitions and critical phenomena , vol . 10 _ ( academic , london , 1986 ) p.1 s. dietrich , in c. domb and j.l . lebowitz ( eds . ) _ phase transitions and critical phenomena , vol . 12 _ ( academic , london , 1988 ) p.1 . m. schick , in j. charvolin , j .- f . joanny , and j. zinn - justin ( eds . ) _ liquids at interfaces _ ( elsevier , amsterdam , 1990 ) p. 415 g. forgacs , r. lipowsky and t.m . nieuwenhuizen , in c. domb and j.l . lebowitz ( eds . ) _ phase transitions and critical phenomana , vol . ( academic , london , 1991 ) chap . 2 d. bonn and d. ross , rep . progr . phys . * 64 * , 1085 ( 2001 ) d.r . clarke , m. rhle , a.p . tomsia ( eds . ) _ low and high temperature wetting : state of the art _ ; _ ann . , vol . 38 _ ( ann . revs . , palo alto , 2008 ) d. bonn , j. eggers , j. indekeu , j. meunier , and e. rolley , rev . phys . * 81 * , 739 ( 2009 ) k. binder , d.p . landau , and m. mller , j. stat . * 110 * , 1411 ( 2003 ) d.b . abraham , phys . lett . * 44 * , 1165 ( 1980 ) m.e . fisher and h. nakanishi , j. chem . phys . * 75 * , 5857 ( 1981 ) h. nakanishi and m.e . fisher , phys . lett . * 49 * , 1565 ( 1982 ) h. nakanishi and m.e . fisher , j. chem . 78 * , 3270 ( 1983 ) k. binder , d.p . landau and d.m . kroll , phys . * 56 * , 2272 ( 1986 ) d.b . abraham and e.r . smith , j. stat . 43 * , 621 ( 1986 ) k. binder and d.p . landau , phys . * , 1745 ( 1988 ) k. binder , d.p . landau , and s. wansleben , phys . rev . b*40 * , 6971 ( 1989 ) e.v . albano , k. binder , d.w . heermann , and w. paul , surf . sci . * 223 * , 151 ( 1989 ) e.v . albano , k. binder , d.w . heermann , and w. paul , j. stat . * 61 * , 161 ( 1990 ) a.o . parry and r. evans , phys . lett . * 64 * , 439(1990 ) a.o . parry and r. evans , physica a * 181 * , 250 ( 1992 ) k. binder , d.p . landau , and a.m. ferrenberg , phys . * 74 * , 298 ( 1995 ) k. binder , d.p . landau , and a.m. ferrenberg , phys . * , 2823 ( 1995 ) k. binder , r. evans , d.p . landau , and a.m. ferrenberg , phys . rev . e*53 * , 5023 ( 1996 ) a. maciolek and j. stecki , phys . rev . b*54 * , 1128 ( 1996 ) a. maciolek , j. phys . a*29 * , 3837 ( 1996 ) e. carlon and a. drzewinski , phys . e*57 * , 2626 ( 1998 ) e.v . albano , k. binder , and w. paul , j. phys.:condens . matter * 12 * , 2701 ( 2000 ) a. de virgiliis , e.v . albano , m. mller , and k. binder , physics a*352 * , 477 ( 2005 ) b.j . schulz , k. binder , and m. mller , phys . rev . e*71 * , 046705 ( 2005 ) l. pang , d.p . landau , and k. binder , phys . * 106 * , 236102 ( 2011 ) l.d . gelb , k.e . gubbins , r. radhakrishnan , and m. slivinska - bartkowiak , rep . phys . * 63 * , 1573 ( 1999 ) e.v . albano , k. binder , d.w . heermann , and w. paul , j. chem . phys . * 91 * , 3700 ( 1989 ) k. binder and d.p . landau , j. chem . 96 * , 1444 ( 1992 ) o. dillmann , w. janke , m. mller , and k. binder , j. chem . phys . * 114 * , 5823 ( 2001 ) k. binder and p.c . hohenberg , phys . b*6 * , 3461 ( 1972 ) k. binder and p.c . hohenberg , phys . rev . b*9 * , 2194 ( 1974 ) k. binder , in c. domb and j.l . lebowitz ( eds . ) _ phase transitions and critical phenomena , vol . 8 _ ( academic , london , 1983 ) p.1 . l. onsager , phys . rev . * 65 * , 117 ( 1944 ) m. blume , phys . rev . * 141 * , 517 ( 1966 ) h.w . capel , physica * 32 * , 966 ( 1966 ) s. sarbach and i.d . lawrie , in c. domb and j.l . lebowitz ( eds . ) _ phase transitions and critical phenomena , vol . 9 _ ( academic , london , 1984 ) p.1 k. binder and d.p . landau , surf . sci . * 61 * , 577 ( 1976 ) j.l . cardy , in c. domb and j.l . lebowitz ( eds . ) _ phase transitions and critical phenomena , vo . ( academic , london , 1987 ) p.55 r.b . pearson , phys . rev . b*22 * , 2579 ( 1980 ) b. nienhuis , j. phys . a : math . gen . * 15 * , 199 ( 1982 ) h.e . stanley , _ an introduction to phase transitions and critical phenomena _ ( oxford univ . press , oxford , 1971 ) k. binder , in d.g . pettifor ( ed . ) _ cohesion and structure of surfaces _ ( elsevier , amsterdam , 1995 ) p. 121 . w. selke and j. yeomans , j. phys . a*16 * , 2789 ( 1983 ) w. selke , d.a . huse , and d.m . kroll , j. phys . a*17 * , 3019 ( 1984 ) w. selke , surf . 144 * , 176 ( 1984 ) r. lipowsky , phys . b*32 * , 1731 ( 1985 ) d.m . kroll , r. lipowsky , and r.k.p . zia , phys . rev . b*32 * , 1862 ( 1985 ) a.o . parry , j.m . romero - enrique and a. lazarides , phys . lett . * 93 * , 086104 ( 2004 ) a.o . parry , c. rascon , n.r . bernardino , and j.m . romero - enrique , j. phys . : condens . matter * 18 * , 6433 ( 2006 ) ; _ ibid _ * 19 * , 416105 ( 2007 ) a.o . parry , c. rascon , n.r . bernardino , and j.m . romero - enrique , phys . * 100 * , 136105 ( 2008 ) a.o . parry , j.m . romero - enrique , n.r . bernardiono , and j. rascon , j. phys . : condens . matter * 20 * , 505102 ( 2008 ) k. binder and j.s . wang , j. stat . phys . * 55 * , 87 ( 1989 ) k. binder , in v. privman ( ed . ) _ finite size scaling and numerical simulation of statistical systems _ ( world scientific , singapore , 1990 ) p. 173 bhattacharjee and j.f . a , * 31 * , 3199 ( 1985 ) s.m . bhattacharjee and j.j . a , * 44 * , 6202 ( 1991 ) k. binder , phys . lett . * 47 * , 693 ( 1981 ) k. binder , rep . phys . * 60 * , 487 ( 1997 ) l.d.c . jaubert , j.t . chalker , p.c.w . holdsworth , and r. moessner . lett . * 105 * , 087201 ( 2010 ) a. milchev , m. mller , k. binder , and d.p . landau , phys . * 90 * , 136101 ( 2003 ) a. milchev , m. mller , k. binder , and d.p . landau , phys . rev . e*68 * , 031601 ( 2003 ) a. milchev , m. mller , k. binder , europhys . 70 * , 348 ( 2005 ) a. milchev , m. mller , and k. binder , phys . rev . e*72 * , 031603 ( 2005 ) a. devirgiliis , e.v . albano , m. mller , and k. binder , j. phys . : condens . matter * 17 * , 4579 ( 2005 ) a. winkler , d. wilms , p. virnau , and k. binder , j. chem . phys . * 133 * , 164702 ( 2010 )

the wetting transition of the blume - capel model is studied by a finite - size scaling analysis of @xmath0 lattices where competing boundary fields @xmath1 act on the first row or last row of the @xmath2 rows in the strip , respectively . we show that using the appropriate anisotropic version of finite size scaling , critical wetting in @xmath3 is equivalent to a `` bulk '' critical phenomenon with exponents @xmath4 , @xmath5 , and @xmath6 . these concepts are also verified for the ising model . for the blume - capel model it is found that the field strength @xmath7 where critical wetting occurs goes to zero when the bulk second - order transition is approached , while @xmath8 stays nonzero in the region where in the bulk a first - order transition from the ordered phase , with nonzero spontaneous magnetization , to the disordered phase occurs . interfaces between coexisting phases then show interfacial enrichment of a layer of the disordered phase which exhibits in the second order case a finite thickness only . a tentative discussion of the scaling behavior of the wetting phase diagram near the tricritical point also is given . pacs : 68.08.bc , 68.35.rh , 64.60f- , 05.70.fh

the total yields and inclusive spectra of different species of hadrons in leptonuclear interactions reflect the complicated space - time structure of the quark string fragmentation , the hadron formation and the secondary intranuclear interactions . the experimental investigations on this topic concern mainly stable hadrons ( pions , kaons , ( anti)protons ) . meanwhile , the data concerning hadronic resonances ( being , predominantly , direct products of the quark string fragmentation ) can provide a valuable additional information about the nuclear medium influence on the leptoproduction processes . hitherto , no data are available for the a- dependence of the total yields and inclusive spectra of hadronic resonances in lepton - induced reactions . + this work is devoted to the study , for the first time , of the a- dependence of @xmath1 meson neutrinoproduction on nuclei . to this end , the data from the skat bubble chamber are used . in section 2 , the experimental procedure is described . the experimental data on the a- dependence of the total yield and inclusive spectra of @xmath1 are presented in section 3 and discussed in section 4 . the results are summarized in section 5 . the experiment was performed with skat bubble chamber @xcite , exposed to a wideband neutrino beam obtained with a 70 gev primary protons from the serpukhov accelerator . the chamber was filled with a propane - freon mixture containing 87 vol% propane ( @xmath8 ) and 13 vol% freon ( @xmath9 ) with the percentage of nuclei h : c : f : br = 67.9:26.8:4.0:1.3 % . a 20 kg uniform magnetic field was provided within the operating chamber volume . + charged current interactions containing a negative muon with momentum @xmath100.5 gev / c were selected . other negatively charged particles were considered to be @xmath11 mesons . protons with momentum below 0.6 gev@xmath12 and a fraction of protons with momentum 0.6 - 0.85 gev@xmath12 were identified by their stopping in the chamber . non - identified positively charged particles were considered to be @xmath13 mesons . events in which errors in measuring the momenta of all charged secondaries and photons were less than 60% and 100% , respectively , were selected . each event is given a weight which corrects for the fraction of events excluded due to improperly reconstruction . more details concerning the experimental procedure , in particular , the reconstruction of the neutrino energy @xmath14 can be found in our previous publications @xcite . + the events with @xmath15 30 gev were accepted , provided that the reconstructed mass @xmath16 of the hadronic system exceeds 2 gev . no restriction was imposed on the transfer momentum squared @xmath17 . the number of accepted events was 4353 ( 5508 weighted events ) . the mean values of the kinematical variables were @xmath18 = 10.7 gev , @xmath19 = 3.0 gev , @xmath20 = 9.6 gev@xmath21 , @xmath22 = 2.8 ( gev / c)@xmath21 . + further , the whole event sample was subdivided , using several topological and kinematical criteria @xcite , into three subsamples : the cascade subsample @xmath23 with a sign of intranuclear secondary interaction , the quasiproton ( @xmath24 ) and quasineutron ( @xmath25 ) subsamples . about 40% of subsample @xmath24 is contributed by interactions with free hydrogen . weighting the quasiproton events with a factor of 0.6 , one can compose a pure nuclear subsample @xmath26 and a quasinucleon subsample @xmath27 . it has been verified @xcite , that the multiplicity and spectral characteristics of secondary particles in the @xmath28 subsample are in satisfactory agreement with those measured with a pure proton ( deuteron ) target . the effective atomic weight corresponding to the subsample @xmath29 is estimated @xcite to be approximately equal to @xmath30 , when taking into account the probability of secondary intranuclear interactions in the composite target . + in order to extract the a- dependence of the @xmath1 neutrinoproduction , we use in the next section the data obtained for subsamples @xmath31 and @xmath29 , as well as the published data @xcite on neutrino - freon interactions ( with @xmath32 ) . the @xmath33 effective mass distribution for subsamples @xmath31 and @xmath29 is plotted in fig . clear signals near the @xmath1 mass , as well as faint signals near the @xmath34 mass are seen . the distributions are fitted by expression @xmath35 where the mass dependence of the background distribution is parametrized as @xmath36 while for @xmath37 and @xmath38 the corresponding relativistic breit - wigner functions @xcite for @xmath1 and @xmath34 were used , taking into account the experimental mass resolution @xmath39 = 35 mev . the mass and width of resonances are fixed as : @xmath40 = 775 mev and @xmath41 = 150 mev from the pdg data @xcite and @xmath42 = 963 mev and @xmath43 = 35 mev from the recent nomad measurements @xcite . + the resulting total yields of @xmath1 and @xmath34 are presented in table 1 , where the skat data for @xmath1 in neutrino - freon interactions @xcite are also given . the data on @xmath34 are corrected for the undetectable mode @xmath44 . for comparison , the data on the @xmath11 yields are shown too . .the mean multiplicities of @xmath1 , @xmath34 and @xmath11 and the ratio @xmath45 . [ cols="<,^,^,^,^",options="header " , ] table 2 presents the yields of @xmath1 and @xmath11 and their ratio in the forward ( @xmath46 , @xmath3 being the feynman variable ) and backward ( @xmath47 ) hemispheres in the hadronic c.m.s . the data on @xmath11 indicate , in accordance with our previous studies ( see details in @xcite ) a clear depletion in the forward hemisphere ( due to the nuclear attenuation ) and a significant enhancement in the backward hemisphere ( due to the secondary inelastic intranuclear interactions ) . a faint indication on similar effects are also seen for @xmath1 . + as it is seen from table 2 and fig . 3 where the distributions on @xmath3 for @xmath1 and @xmath11 are plotted , the overwhelming part of @xmath1 in @xmath48 interactions and the most part of that in nuclear interactions are produced in the forward hemisphere , as a result of the current quark ( @xmath49 or @xmath50 ) fragmentation into a favorable hadron @xmath1 ( which can contain the current quark ) , while the production of the ( unfavorable ) @xmath11 meson occurs mainly in the central region . as a result , the ratio @xmath51 for the fastest hadrons ( with @xmath52 ) significantly exceeds that in other ranges of the variable @xmath3 , as it can be seen from fig . 4 . note , that a comparison of our data with those obtained at higher energies does not reveal any @xmath16- dependence of the @xmath51 ratio in the forward hemisphere . indeed , the values of @xmath53 = [email protected] at @xmath55 0.5 for @xmath48 - interactions and @xmath56 = [email protected] at @xmath55 0 for @xmath57 - interactions obtained in this work are compatible , respectively , with @xmath58 = [email protected] measured in @xmath59 - interactions at @xmath19 = 6.1 gev @xcite and @xmath60 = [email protected] measured in @xmath61 - interactions at @xmath19 = 4.8 gev @xcite . + it is interesting to compare the relative yield of strange and non - strange favorable mesons , @xmath62 and @xmath1 , with that of unfavorable ones , @xmath63 and @xmath11 . the former can be extracted using our recent data on @xmath62 neutrinoproduction @xcite . the ratio of the @xmath62 and @xmath1 total yields is estimated to be [email protected] ( for @xmath0 = 1 ) and [email protected] ( for @xmath64 21 ) which seems to exceed that for @xmath63 and @xmath11 measured in @xcite : [email protected] ( for @xmath0 = 1 ) and [email protected] ( for @xmath64 21 ) . these data , therefore , indicate , that the strangeness content in favorable mesons in neutrino - induced reactions is higher than that for unfavorable ones . in figs . 5 - 7 , the inclusive spectra of @xmath1 for quasinucleon and nuclear interactions are compared with those measured in neutrino - freon interactions @xcite . 5 shows the invariant distribution on @xmath3 . it is seen , that the data do not exhibit a significant @xmath0- dependence in the forward hemisphere , while at @xmath65 0 the @xmath1 yield in @xmath57 - interactions is enhanced as compared to @xmath48 - interactions . + the distributions on the variable @xmath66 are presented in fig . it is seen , that they tend to shift towards lower values of @xmath4 with increasing a , as expected due to the effects of secondary intranuclear interactions . 7 shows the @xmath4- dependence of the ratio @xmath67 . the data exhibit no significant dependence on @xmath0 , thus indicating that the nuclear effects are approximately the same for @xmath1 and pions ( averaged over @xmath68 and @xmath11 ) . the data presented in the previous section indicate on a small but not negligible role of nuclear effects in the @xmath1 neutrinoproduction . it is interesting to clarify whether these effects are compatible with expectations based on the accounting for intranuclear interactions of secondary pions , @xmath69 , resulting in a @xmath1 multiplicity gain @xmath70 ( mainly at @xmath47 ) , as well as the absorption processes , @xmath71 ( no @xmath1 ) , resulting in a yield reduction , @xmath72 ( mainly at @xmath46 ) . the details of a simple model incorporating these processes can be found in @xcite and references therein . + the gain @xmath73 induced by pions with momenta @xmath74 is determined by the differential multiplicity of pions , @xmath75 , the mean probability of their secondary inelastic interactions @xmath76 averaged over the nuclei of the composite target , and by the mean multiplicity @xmath77 of @xmath1 in inelastic @xmath78 interactions . the values of @xmath77 extracted ( with an uncertainty of about @xmath54 15% ) from the available experimental data @xcite vary from 0.002 at @xmath79 = 0.9 - 1 gev@xmath12 up to 0.22 at @xmath79 = 10 - 15 gev@xmath12 ( the end of the pion spectrum in this experiment ) . + the product @xmath80 was integrated over the momentum spectra of charged pions measured in quasinucleon interactions . the contribution from secondary interactions of @xmath81 mesons is assumed to be the average of those for @xmath68 and @xmath11 mesons . the resulting values of @xmath82 are found to be [email protected] for @xmath83 45 and 0.023@xmath84 for @xmath83 21 . the latter value can be compared with the experimental value of @xmath85 = [email protected] , inferred from the last two lines of table 2 ( note , that in the evolution of the error in @xmath86 the correlation between the values of @xmath87 and @xmath88 was taken into account here and below ) . + in order to estimate the @xmath1 yield decreasing , @xmath89 , one needs to know the @xmath1 absorption cross section @xmath90 via the channel @xmath71 ( no @xmath1 ) . as @xmath90 is not known , we use tentative values in between @xmath91 mb . at these values , the probability @xmath92 of the @xmath1 absorption is estimated to be 23@xmath546% for @xmath83 45 and 18@xmath545% for @xmath83 21 ( see @xcite for details of the absorption probability calculations ) . with these probabilities , the value of @xmath93 is equal to [email protected] and [email protected] , respectively . the latter value can be compared with @xmath94 = [email protected] inferred from the first two lines of table 2 . + the total multiplicity gain , @xmath95 , is predicted to be @xmath96 = [email protected] for @xmath83 21 and [email protected] for @xmath83 45 . these values are compatible with a rather weak variation of @xmath97 with @xmath0 ( cf . table 1 ) , resulting in a total multiplicity gain @xmath98 = [email protected] for @xmath83 21 and [email protected] for @xmath83 45 . the large errors in @xmath98 do not allow to check quantitatively the predictions of the model incorporating secondary intranuclear interactions . new experimental data on @xmath1 production in @xmath48 and @xmath99 ( @xmath100 ) interactions are obtained at intermediate energies ( @xmath18 = 10.7 gev , @xmath19 = 3.0 gev ) . for the first time , nuclear effects in @xmath1 neutrinoproductions are observed . these effects for the total yield @xmath97 of @xmath1 are found to be rather small and compatible with those for pions . using also the skat data for @xmath83 45 , the slope parameter @xmath101 in the exponential parametrization of the total yields ( @xmath102 ) is found @xmath103 which agrees with that for @xmath11 mesons , @xmath104 . the ratio of @xmath105 is found to be independent of @xmath0 being equal about 0.1 . a comparison with the data at higher energies reveals a slight increase of this ratio with increasing energy , while no energy dependence is observed for that in the forward hemisphere ( @xmath46 ) . + a comparison of the @xmath1 inclusive spectra in @xmath48 and @xmath57- interactions indicates , that the major influence of the nuclear medium consists in their shifting towards smaller values of @xmath3 and @xmath4 , thus indicating on a non - negligible role of the secondary intranuclear interactions . the predictions of a simplified model incorporating the latter are found to be in qualitative agreement with experimental data . + a comparison of the relative yield of the favorable strange and non - strange mesons ( @xmath62 and @xmath1 ) with that for unfavorable ones ( @xmath63 and @xmath11 ) indicates , that the strangeness content in the former is higher than in the latter . + * acknowledgement . * the activity of one of the autors ( h.g . ) is supported by cooperation agreement between desy and yerphi signed on december 6 , 2002 . the autors from yerphi acknowledge the supporting grants of calouste gulbenkian foundation and swiss fonds `` kidagan '' . 00 = 0.pt = 0.pt = 0.pt v.v.ammosov et al . chastits at . yadra * 23 * , 648 , 1992 [ sov . j. part . * 23 * , 283 , ( 1992 ) ] . n.m.agababyan et al . , ( skat coll ) , yerphi preprint n 1535(9 ) ( yerevan , 1999 ) . n.m.agababyan et al . , ( skat coll ) , yad . fiz . * 66 * , 1350 ( 2003 ) . [ phys . of at . nucl . * 66 * , 1310 ( 2003 ) ] . n.m.agababyan et al . , ( skat coll),yerphi preprint n 1578(3 ) ( yerevan , 2002 ) . n.m.agababyan et al . , ( skat coll ) , yad . * 68 * , 1241 ( 2005 ) . n.m.agababyan et al . , ( skat coll),yerphi preprint n 1597(1 ) ( yerevan , 2005 ) ; hep - ex/0504024 . v.v.ammosov et al . , ( skat coll . ) , yad . fiz . * 46 * , 131 ( 1987 ) . * 46 * , 80 ( 1987 ) ] . j.d.jackson , nuovo cim . * 34 * , 1644 ( 1964 ) . review of particle physics , phys . b*592 * , 39 ( 2004 ) . p.astier et al . , ( nomad coll . ) , nucl . b*601 * , 3 ( 2001 ) . w.wittek et al . , ( bebc wa59 coll . ) , z. phys . c*44 * , 175 ( 1989 ) . g.t.jones et al . , ( wa21 coll . ) , z. phys . c*51 * , 11 ( 1991 ) . n.m.agababyan et al . , ( skat coll ) , yerphi preprint n 1548(2 ) ( yerevan , 2005 ) ; hep - ex/0504040 . v.flaminio et al . , compilation cern - hera 83 - 01 , 1983 .

the @xmath0- dependence of @xmath1 meson production in neutrino - induced reactions is investigated for the first time , using the data obtained with skat bubble chamber . the nuclear medium influence on the @xmath1 total yield and inclusive distributions ( on @xmath2 and feynman @xmath3 variables ) is found to be approximately the same as for pions . it is shown , that these distributions , with increasing @xmath0 , tend to shift toward smaller values of @xmath4 and @xmath3 , thus indicating on an increasing role of secondary intranuclear interactions . the predictions of a simplified model incorporating the latter are found to be in qualitative agreement with experimental data . 1.0 cm skat collaboration n.m . agababyan@xmath5 , v.v . ammosov@xmath6 , m. atayan@xmath7 , + n. grigoryan@xmath7 , h. gulkanyan@xmath7 , a.a . ivanilov@xmath6 , + zh . karamyan@xmath7 , v.a . korotkov@xmath6 @xmath5 joint institute for nuclear research , dubna , russia @xmath6 institute for high energy physics , protvino , russia @xmath7 yerevan physics institute , armenia * yerevan 2005 *

large complex systems are composed of various interconnected components . the measure of the behavior of a single component thus results from the superimposition of different factors acting at different levels . common factors such as global trends or external socio - economical conditions obviously play a role but usually different sub - units ( such as users in the internet , states or regions in a country ) will react in different ways and add their local dynamics to the collective pattern . for example , the number of downloads on a website depends on factors such as the time of the day but one can also observe fluctuations from a user to another one @xcite . in the case of criminality , favorable socio - economical conditions will impose a global decreasing trend while local policies will affect the regional time series . in the case of financial series , the market imposes its own trend and some stocks respond to it more or less dramatically . in all these cases it is important to be able to distinguish if the stocks or regions are at the source of their fluctuations or if on the opposite , they just follow the collective trend . extracting local effects in a collection of time series is thus a crucial problem in assessing the efficiency of local policies and more generally , for the understanding of the causes of fluctuations . this problem is very general and as the availability of data is always increasing particularly in social sciences , it becomes always more important for the modeling @xcite and the understanding of these systems . there is obviously a huge literature on studying stochastic signals @xcite ranging from standard methods to more recents ones such as the detrended fluctuation analysis @xcite , independent component analysis @xcite , and separation of external and internal variables @xcite . most of these methods treat the internal dynamics as a small local perturbation with zero mean which is in contrast with the method proposed here . in a first part we present the method . in a second part , we test it on synthetic series generated by correlated random walkers . we then apply the method to empirical data of crime rates in the us and france , and obesity rates in the us , for which , to our knowledge , no general quantitative method is known to provide such separation between global and local trends . in general , one has a set of time series @xmath2 where @xmath3 and we will assume that the number @xmath4 of units is large . the index @xmath5 refers to a particular unit on a specific scale such as a region , city , a country . the problem we address consists in extracting the collective trend and the effect of local contributions . one way to do so is to assume the signal @xmath6 to be of the form @xmath7 where the ` external ' part , @xmath8 , represents the impact on the region @xmath5 of a global trend , while the ` internal ' part , @xmath9 , represents the contribution due to purely local factors . usually , in order to discuss the impact of local policies , one compares a regional ( local ) curve @xmath10 to the average ( the national average in case of regions of a country ) computed as @xmath11 ( or @xmath12 if one has intensive variables and populations @xmath13 ) . although reasonable at first sight , this assumes that the local component is purely additive : @xmath14 _ local term_. in this article , following @xcite , we will rather consider the possibility of having both multiplicative and additive contributions . more specifically , we assume @xmath15 where @xmath16 is a collective trend common to all series , and which affects each region @xmath5 with a corresponding prefactor @xmath17 . these coefficients are assumed to depend weakly on the period considered , ie . to vary slowly with time . we thus write @xmath18 we first note that the global trend @xmath19 is known up to a multiplicative factor only ( one can not distinguish @xmath20 from @xmath21 whatever @xmath22 ) and we will come back to this issue of scale later . also , the purely additive case is recovered if the @xmath17 s are independent of @xmath5 . if on the contrary the @xmath17 s are different from one region to the other , the national average ( [ eq : fnat ] ) , @xmath23 , is then given by @xmath24 here and in the following we denote the sample average , that is the average over all units @xmath5 , by a bar , @xmath25 , and the temporal average by brackets @xmath26 . the ` naive ' local contribution is then estimated by the difference with the national average @xmath27 the estimated local contribution @xmath28 can thus be very different from the original one , @xmath29 , and the difference @xmath30 will be very large at all times @xmath31 where @xmath16 is large ( note that the conclusion would be the same by taking the national average as @xmath32 ) . this demonstrates that comparing local time series with the naive average could in general be very misleading . beside the correct computation of the external and internal contributions , the existence of both multiplicative and additive local contributions implies that the effect of local policies must be analyzed by considering both how the local unit @xmath5 follows the global trend ( @xmath17 ) and how evolves the purely internal contribution ( @xmath33 ) . in a previous study @xcite , menezes and barabasi proposed a simple method to separate the two contributions , internal ( @xmath9 ) and external ( @xmath34 written as @xmath35 ) . they assume that the temporal average @xmath36 is zero , and compute the external and internal parts by writing @xmath37 and @xmath38 . this method can be shown to be correct in very specific situations , such as the case where @xmath10 is the fluctuating number of random walkers at node @xmath5 in a network , but in many cases however , one can expect that the local contributions have a non zero sample average and the method of @xcite will yield incorrect results . indeed , if the hypothesis eq . ( [ eq : hyp ] ) is exact , this method would give for @xmath19 the estimate @xmath39 , and in the limit @xmath40 for @xmath41 would lead to the estimates @xmath42 and @xmath43 , which are different from the exact results , except if @xmath44 . in order to separate the two contributions we propose in this article a totally different approach , by taking an independent component analysis point of view in which we do not assume that the local contribution has a zero average ( over time and/or over the regions ) . to express the idea that the ` internal ' contribution is by definition what is specifically independent of the global trend , and that the correlations between regions exist essentially only through their dependence in the global trend , we impose that the global trend is statistically independent from local fluctuations @xmath45 ( we denote by @xmath46 the connected correlation @xmath47 ) , and that these local fluctuations are essentially independent from region to region , that is for @xmath48 @xmath49 where this statement will be made more precise below . we show that , for large @xmath4 , these constraints ( [ eq:<wfint > ] ) , ( [ eq:<gigj > ] ) are sufficient to extract estimates of the global trend @xmath19 and of the @xmath17 s . we denote by @xmath50 the average of @xmath19 and by @xmath51 its dispersion , so that we write @xmath52 with @xmath53 and @xmath54 . if we denote by @xmath55 and @xmath56 , we have @xmath57 with @xmath58 note that @xmath59 . if we now consider the correlations between these centered quantities , @xmath60 , we find @xmath61 if we assume that for @xmath48 @xmath62 is negligible ( of order @xmath63 ) compared to @xmath64 ( which is what we mean by having small correlations between internal components , eq . ( [ eq:<gigj > ] ) ) , from this last expression we can show that at the dominant order in @xmath4 , we have @xmath65 these equations lead to @xmath66 which is valid when @xmath67 . we note that our method has a meaning only if strong correlations exist between the different @xmath10 s and if it is not the case , the definition of a global trend makes no sense and the approximation used in our calculations are not valid . in the supporting information ( section si1 ) we show that the factors @xmath68 s can also be computed as the components of the eigenvector corresponding to the largest eigenvalue of @xmath69 - a method which is valid under the weaker assumption of having a small number ( compared to @xmath4 ) of non diagonal terms of the matrix @xmath70 which are not negligible . once the quantities @xmath68 are known , we can compute the global normalized pattern @xmath71 with the reasonable estimator given by @xmath72 , @xmath73 indeed , @xmath74 and since the quantity @xmath75 is a sum of independent variables with zero mean , we can expect it to behave as @xmath76 . we can show that this actually results from the initial assumptions . indeed , by construction @xmath77 and the second moment is @xmath78 by assumption we have @xmath79 if @xmath80 and we thus obtain @xmath81 . the computation of the @xmath68 s and of @xmath82 is equivalent to an independent component analysis ( ica ) @xcite with a _ single _ source ( the global trend ) and a large number @xmath4 of sensors . however , in contrast with the standard ica , we are not interested in getting only the sources ( here the trend @xmath82 ) , but also the internal contributions ( which , in a standard ica framework , would be considered as noise terms , typically assumed to be small ) . we have already the @xmath68 s , and since @xmath71 has been calculated we can compute @xmath83 . we thus obtain at this stage @xmath84 this is a set of @xmath4 equations for @xmath85 unknown ( @xmath86 and the @xmath87 s ) and we are thus left with one free parameter , the ratio @xmath86 . knowing its value would give the @xmath4 local averages , the @xmath36 s . less importantly one may want also to fix the average @xmath50 ( hence both @xmath50 and @xmath51 ) in order to fully determine the pattern @xmath16 : this will be of interest only for making a direct comparison between this pattern and the national average ( [ eq : fnat ] ) . this equation ( [ eq : mui ] ) suggests a statistical linear correlation between @xmath88 and @xmath89 , with a slope given by @xmath86 . we will indeed observe a linear correlation in the data sets ( next section , figure @xmath90 ) however , it could be that the @xmath36 s themselves are correlated with the @xmath68s . hence , and unfortunately , a linear regression can not be used to get an unbiased estimate of the parameter @xmath86 . in the absence of additional information or hypothesis this parameter remains arbitrary . however one may compare the qualitative results obtained for different choices of @xmath86 : which properties are robust , and which ones are fragile . in particular one would like to be able to access how a given region is behaving , compared to another given region , and/or to the global trend . to do so , in the applications below we will in particular analyze : ( i ) the correlations between the two local terms , @xmath68 and @xmath36 ; ( ii ) the robustness of the rank given by the @xmath36 s ; ( iii ) the sign of @xmath87 ; ( iv ) the quantitative and qualitative similarities between @xmath91 and the naive estimate @xmath28 . we will focus on two particular scenarios . first , one may ask the global trend to fall ` right in the middle ' of the @xmath4 series . there are different ways to quantify this . one way to do so is to note that , in the absence of internal contribution , @xmath92 would be equal to @xmath19 , hence @xmath93 would be equal to @xmath86 . therefore we may compute @xmath86 by imposing @xmath94 which is thus equivalent to impose @xmath95 an alternative is to ask the resulting @xmath9 to be as close as possible to the naive ones ( eq . ( [ eq : naive ] ) ) , by minimizing @xmath96 which gives @xmath97 in both cases one may then fix @xmath50 from @xmath98 or by imposing @xmath99 for some arbitrary chosen @xmath100 . finally , one may rather ask for a conservative comparison with the naive approach by minimizing the difference between @xmath19 and @xmath101 : either by writing @xmath98 ( or @xmath99 ) and @xmath102 , or by minimizing @xmath103 , which gives @xmath104 for @xmath4 is large , one can check that the results depend weakly on any one of these reasonable choices . the second scenario considers the correlations between the @xmath87 s and the @xmath68 s . as we will see , the first hypothesis leads to a strictly negative correlation . an alternative is thus to explore the consequences of assuming no correlations , hence asking for @xmath105 which implies that the slope of the observed linear correlation @xmath106 with @xmath68 gives the value of @xmath86 . as explained above , for each application below we will discuss the robustness of the results with respect to these choices of the parameter @xmath86 . we can now summarize our method . it consists in ( i ) estimating the @xmath68 s using eq . ( [ eq : ai ] ) ( or using the eigenvector corresponding to the largest eigenvalue of the correlation matrix , section si1 ) , ( ii ) computing @xmath82 using eq . ( [ eq : w ] ) , and finally ( iii ) comparing the results for different hypothesis on @xmath86 as discussed above . we propose to call this method the _ external trend and internal component analysis _ ( * etica * ) . we note that if the hypothesis eq . ( [ eq : hyp ] ) , ( [ eq:<wfint > ] ) , ( [ eq:<gigj > ] ) are correct , the method gives estimates of @xmath82 , the @xmath68s ( hence of @xmath107 ) which become exact in the limit @xmath31 and @xmath4 large , and a good estimate of the full trend @xmath19 ( hence of the @xmath87 ) whenever this trend , qualitatively , does fall ` in the middle ' of the time series . once we have extracted with this method the local contribution @xmath9 , and the collective pattern @xmath16 together with its redistribution factor @xmath17 for each local series , we can study different quantities , as illustrated below on different applications of the method . in general , although this method gives a pattern @xmath16 very similar to the sample average @xmath108 , we will see that there is non trivial structure in the prefactors @xmath17 s leading to non trivial local contributions @xmath91 . in some cases one may expect to have , in addition to the local contribution , a linear combination of several global trends ( a small number of sources ) : we leave for future work the extension of our method to several external trends . we first test our method on synthetic series and we then illustrate it on crime rate series ( in the united states and in france ) and on us obesity rate series . for the crime rates , a plot of the time series shows that obviously a common trend exists ( fig . c + after computing the internal and external terms , we perform different tests in order to assess the validity of the approach . in particular , figure @xmath90 shows a plot of the local factors @xmath68s versus the data time - averages , the @xmath106 s . c + one observes a statistical linear correlation in the four set of time series . we stress that the @xmath68 s are computed from the covariance matrix of the data , hence after removing the means from the time series . the fact that we do observe a linear correlation is thus a hint that our hypothesis on the data structure is reasonable ( in contrast the very good linear correlation observed in @xcite can be shown to be an artefact of the method used in these works , leading to an exact proportionality independently of the data structure , ( see the section si2 ) . we now discuss in more detail the synthetic series , each one of the crime rate data sets , and the obesity rate . we can illustrate our method on the case of correlated random walkers described by the equation @xmath109 where @xmath110 is the global trend imposed to all walkers and the @xmath111 are gaussian noises but with possible correlations between different walkers @xmath112/12 $ ] where @xmath113 and @xmath114 are tunable parameters ( see the supporting section si3 ) . for @xmath115 , the random noises @xmath111 are independent and our method is very accurate : we choose for example a sinusoidal trend @xmath116 and we plot in the figure @xmath117 the original signal , the exact local contribution and the local contribution computed with our method . c + when the correlation between walkers is increasing we study the pearson correlation coefficient between the original local contribution and the estimate provided by our method , and we observe that our method is indeed accurate as long as the correlations between the @xmath118 s are not too large , which corresponds here to the condition @xmath119 . in criminology an essential question concerns the impact of local policies , a subject of much debate @xcite . in order to assess these local effects ( at the level of a state or a region ) , most authors consider the difference of a state evolution with the national average . as we noticed above this may lead to incorrect predictions . in this second part of applications , we thus illustrate our method on the analysis of the series of crime rates in @xmath120 states in the us @xcite for the period @xmath121 , and about @xmath122 _ dpartements _ of france @xcite for the period @xmath123 . on fig . @xmath124 we represent these time series normalized by their time average . the observed data collapse confirms the existence of a collective pattern ( we also show on this plot the collective pattern @xmath16 obtained with our method ) . for the french case , we have withdrawn outliers which do not satisfy our initial assumptions . the series of these _ dpartements _ are indeed uncorrelated with the rest of crime rates and can not be incorporated in the calculation of the collective pattern . we apply our method to these data and extract @xmath16 , the @xmath68 s and @xmath91 . as already mentioned , we plot on fig . @xmath90 the @xmath68 s vs. the averages @xmath106 , exhibiting a statistical linear correlation . we can check a posteriori that all conditions assumed in our calculation are fulfilled ( zero @xmath125 and small @xmath126 , see si1 ) . also , we checked that the coefficients @xmath17 do not vary too much the period considered , which is an important condition for our method ( see the discussion on different datasets in the si4 ) . in order to assess quantitatively the importance of local versus external fluctuations , we study in particular the ratio of dispersions defined by @xmath127 where the external contribution is the standard deviation of @xmath128 , that is @xmath129 , and the internal one is given by @xmath130 . note that these quantities @xmath131 , being based on fluctuations , does not depend on @xmath132 . this quantity is found in both cases in france and in the us larger than one . this indicates that external factors always dominate over local fluctuations , while local policies seem to play a minor role . in the case of crime , these external effects might be socio - economical factors such as unemployment , density , etc . in addition to compute the average of the @xmath131 s , we can also observe the time evolution of the heterogeneity defined by the sample variances of the different components . we first observe on fig . @xmath133 that large fluctuations are observed in the transition period of mid-@xmath0 s during which crime rates increased significantly . c + we also observe for the usa that until @xmath134 , fluctuations were essentially governed by local effects but that this trend is inverted and increases in the period post-@xmath1 s . in particular during the period @xmath135 during which one observes a decline of crime rates @xcite , it is the collective trend which determines the fluctuations . even we have presented results for reasonable choices of the parameter @xmath51 ( in the following we make the harmless choice @xmath136 ) , one can ask the question of the robustness of different observed properties . first , we can compare the predictions for @xmath51 obtained for the different assumptions used in this paper . in the upper panels for figs . @xmath137 and @xmath138 we show for the us ( france ) , the quantities @xmath139 , @xmath140 and @xmath141 . c + we see on these figures that these quantities are zero for values of @xmath51 which are very close . we also compute the fraction of time @xmath142 for which @xmath91 and the naive calculation @xmath143 have different signs . c + we plot in the lower panels of figs . @xmath133 and @xmath137 , the quantity @xmath144 showing a that for this range of @xmath51 , the signs of @xmath36 and @xmath143 are the same for about @xmath145 of the time period . we can also study the sign @xmath87 versus @xmath51 and we can observe some robustness . in particular , in the us case , approximately @xmath138 states ( ca , nv , mo , mi , ny , az ) have a positive local contribution ( in the range @xmath146 $ ] while @xmath138 states have always a negative local contribution ( vt , ga , la , nh , ct , ms ) . in these cases we can reasonably imagine that local policies have a noticeable effect . finally , we can also analyze the ranking of the local contributions @xmath36 versus @xmath51 by studying kendall s @xmath147 for the two consecutive series @xmath148 and @xmath149 . in both cases ( france and us ) we observe a @xmath147 larger than @xmath150 for the range chose @xmath151 $ ] ( the control case for a random permutation being less than @xmath152 ) indicating a large robustness of the ranking . this means that independently of the assumption used to compute @xmath51 we can rank the different regions according to the importance of their local contribution . the prevalence of obesity ( defined as a body max index - bmi , which is the ratio of the body mass to the square of the height - larger than @xmath153 ) is rapidly increasing in the world @xcite and reached epidemic proportion in the us and is now a major public health concern @xcite . disparities by sex and between ethnic groups have been observed in the prevalence of obesity @xcite , but few studies focus on the effect of local factors and policies on the obesity rate . we thus apply our method to data from the cdc @xcite which describe the percentage of the population which is obese for each states in the us and for the period 1995 - 2008 . as in the crime rate case , we can compare the variances for the internal and external contributions ( see si5 ) and we observe that the external contribution is dominating since the year @xmath154 . this result means that the global trend is the major cause of the evolution of obesity in different states . we can get more detailed information about the specific behavior of the states by studying the ratio @xmath131 defined in eq . ( 26 ) and the ratio of the fraction of the time average local contribution to the total signal @xmath155 . we represent these two quantities in a plane ( see figure @xmath156 ) and we first note that for all states @xmath157 which means that fluctuations are mainly governed by the global trend . c + we can also divide the states into two groups ( with @xmath158 and @xmath159 ) . for large and positive @xmath160 , the states have a small @xmath17 which means that these states are the less susceptible to the global trend , while in the opposite case , the states are governed by the global trend . within each group we can then distinguish the states according to their level of fluctuations ( @xmath131 close to or much larger than one ) . the states arizona , georgia , and oklahoma for example have very little local contribution and their variations is dominated by the global trend . in this respect , states such as dc , indiana are very different from the first group . more generally , we can see on this figure that states with large prevalence display very different values of @xmath161 . this result points toward the fact that describing states by their prevalence only can be very misleading and can hide important dynamical behaviors . finally , we also computed the quantities @xmath160 and @xmath131 using the ` naive ' local contribution using the national average @xmath162 defined in eq . ( [ eq : fnat ] ) by @xmath163 . we represent in figure @xmath164 the difference as vectors of components given by @xmath165 and we can see on this figure that for roughly half of the states the naive calculation of the local contribution can be very misleading . in this article we adressed the crucial problem of extracting the local components of a system governed by a global trend . in this case , comparing the local signal to the average is very misleading and can lead to wrong conclusions . we applied this method to the example of crime rates series in the us and france and our analysis revealed surprising facts . the important result is about the importance of fluctuations which after the @xmath1 s in the us are governed by external factors . this result suggest that understanding the evolution of crime rates relies mostly on the identification of global socio - economical behavior and not on local effects such as state policies etc . in particular , this result could also help in understanding the decreasing trend observed in the us and which so far remains a puzzle @xcite . in the case of obesity , we show that since the year @xmath154 , external factors dominate , and maybe more importantly that states with the same level of prevalence have very different dynamical behaviors , thus calling for the need of a detailled study state by state . however one may expect an even better signal analysis by assuming that there are several independent external trends : it will be interesting to see if our approach , combined with the more standard ica techniques , can be generalized to the case of several global trends ( a small number of sources ) . the recent availability of large amounts of data in social systems call for the need of tools able to analyze them and to extract meaningful information and we hope that our present contribution will help in the understanding of these systems where the local dynamics is superimposed to collective trends . * acknowledgements * : we thank the anonymous referees for constructive remarks , in particular about the applicability conditions of our method . this work is part of the project `` dyxi '' supported by the french national research agency , the anr ( grant anr-08-sysc-008 ) . 1 huberman , b.a . ( 2001 ) _ the laws of the web _ ( mit press , cambridge ) . castellano , c. , fortunato , s. , loreto , v. ( 2009 ) statistical physics of social dynamics . _ reviews of modern physics _ , 81 , 591 - 646 . kautz , h. , schreiber , t. ( 1997 ) _ nonlinear time series analysis _ ( cambridge university press , cambridge ) . peng . et al . ( 1994 ) mosaic organization of dna nucleotides _ physical review e _ 49:1685 - 1689 . comon p. ( 1994 ) independent component analysis : a new concept ? _ signal processing _ , 36(3):287314 ; cardoso , j .- f . ( 1997 ) statistical principles of source separation , _ proc . of sysid97 ( 11th ifac symposium on system identification , fukuoka , japan ) _ pp . 1837 - 1844 ; hyvrinen a. , j. karhunen j. and oja e. ( 2001 ) _ independent component analysis _ , new york : wiley . de menezes , m.a . , barabasi , a .- ( 2004 ) fluctuations in network dynamics . . de menezes , m.a . , barabasi , a .- ( 2004 ) separating internal and external dynamics of complex systems . _ 93:068701 . james , p.t . , leach , r. , kalamara , e. , shayeghi , m. ( 2001 ) the worldwide obesity epidemic . _ obesity research _ , 9:228s-233s . mokdad et al . ( 1999 ) the spread of obesity epidemic in the united states , 1991 - 1998 . _ journal of the american medical association _ 282 : 1519 - 1522 . ogden , c.l . , et al . ( 2006 ) prevalence of overweight and obesity in the united states , 1999 - 2004 . _ journal of the american medical association _ , 295:1549 - 1555 . hedley , a.a . , et al . ( 2004 ) prevalence of overweight and obesity among us children , adolescents , and adults , 1999 - 2002 . _ journal of the american medical association _ , 291:2847 - 2850 . wang , y. , beydoun , m.a . ( 2007 ) the obesity epidemic in the united states gender , age , socioeconomic , racial / ethnic , and geographic characteristics : a systematic review and meta - regression analysis . _ epidemiol rev _ , 29:628 . center for disease control and prevention http://apps.nccd.cdc.gov / brfss/. de maillard j. and roch s. ( 2004 ) crime and justice in france : time trends , policies and political debate _ european journal of criminology _ 1:111 - 151 zimring , f. ( 2007 ) _ the great american crime decline _ ( oxford university press , oxford ) . united states : uniform crime report state statistics from @xmath166 http://www.disastercenter.com/crime/ france : institut national des hautes etudes de scurit ( http://www.inhes.interieur.gouv.fr/bulletin-annuel-112.html ) ; la documentation franaise : criminalit et dlinquance constates en france - tome i : donnes gnrales , nationales , rgionales et dpartementales ( http://www.ladocumentationfrancaise.fr/rapports-publics/084000201/ ) levitt , s.d . ( 2004 ) understanding why crime fell in the 1990s : four factors that explain the decline and six that do not . _ journal of economic perspectives _ 18:163 - 190 . the data correlation matrix @xmath69 is known to provide useful information , in particular for the analysis of financial time series @xmath167 $ ] or in other fields , e.g. in protein structure analysis @xmath168 $ ] . the first , largest , eigenvalue is related to a global trend , and usually one is interested in the small number of intermediate eigenvalues : the associated eigenvectors give the relevant correlations in the data e.g. allows to extract the sectors in financial time series . here , making explicit use of our hypotheses , we extract from the first eigenvector of the correlation matrix the @xmath68 factors which give how the global trend is amplified or reduced at the local level . we have @xmath169 where @xmath170 . if @xmath171 is a normalized eigenvector ( @xmath172 ) of @xmath173 with eigenvalue @xmath174 : @xmath175 , we have @xmath176 we can have @xmath177 which implies that @xmath171 is also eigenvector for @xmath178 which in general is unlikely ( there are no reasons that eigenvectors of @xmath178 are orthogonal to @xmath179 ) . if @xmath180 we then obtain @xmath181 and @xmath182 for the largest eigenvalue , we will neglect at first order the second term of the rhs of this last equation , which leads to @xmath183 . since @xmath171 is normalized , we obtain @xmath184 this approximation is justified if @xmath185 is small compared to @xmath186 and thus @xmath187 since @xmath188 , this approximation is justified if @xmath189 is of order @xmath4 and not of order @xmath190 . this is correct if @xmath178 is diagonal ( which means that the external components are not correlated @xmath191 ) , but also if the number of non - zero terms of @xmath192 is finite compared to @xmath4 , or in other words if @xmath178 is a sparse matrix . we compared the values of @xmath68 computed with the method exposed in the text and with the eigenvector method . results are reported in the figures ( 9,10,11 ) . c + [ fig : eigen1 ] c + [ fig : eigen2 ] c + [ fig : eigen3 ] we see that indeed for the crime rates in the us and in france , @xmath192 is indeed negligible which demonstrate that the correlations of the internal contributions between different states in the us are negligible . this is not the case for the stocks in the @xmath193 where we can observe ( small ) discrepancies between the two methods , a result which supports the idea of sectors in the @xmath193 . we show that the scaling @xmath194 observed by de menezes and barabasi in @xmath195 $ ] is actually built in the method proposed by these authors : it is a direct consequence of their definitions of the internal and external parts , and it does not depend on the data structure . indeed , let @xmath196 be an arbitrary data set such that @xmath197 . for @xmath198 , following @xmath199 $ ] define @xmath200 by @xmath201 and @xmath202 by @xmath203 then , from these definitions and without any hypothesis or constraint on the data other than @xmath197 , one has @xmath204 and @xmath205 hence @xmath206 with @xmath207 hence , one has always @xmath208 the dispersion of the external component , if defined from ( [ eq : aib ] ) and ( [ eq : fextb ] ) , is thus _ exactly _ proportionnal to the mean value of the local data . we considered the case where the external trend is @xmath209 the gaussian noises are given by @xmath210 where the @xmath211 and @xmath212 are independent , uniform random variable of zero mean and variance equal to @xmath213 . in this case , the correlation between different noises are governed by the parameters @xmath113 and @xmath114 @xmath214 when @xmath115 , the variables @xmath215 and @xmath216 are independent ( for @xmath80 ) and we can monitor the correlations by increasing the value of @xmath114 . we plot in figure @xmath217 , @xmath218 random walkers in the usual uncorrelated case and in presence of correlations . c + [ fig : examplerw ] in this simple case the exact result is given by @xmath219 , @xmath220 , and @xmath221 . the important condition for the validity of the method is given by @xmath222 and is given here by @xmath223 for @xmath115 , the random noises are independent and our method is very accurate as shown in the main text . more generally , in order to assess quantitatively the efficiency of the method , we compute the pearson correlation coefficient between the exact @xmath29 and the estimate @xmath224 computed with the method . we plot in figure @xmath225 this coefficient versus @xmath226 . c [ fig : pearson ] this figure confirms the fact that our method is valid and very precise provided that the correlations between local contributions are not too large ( here @xmath227 ) . we can compute the quantities @xmath17 for the interval @xmath228 $ ] and by letting @xmath31 vary . we then obtain for the crime in the us ( in the case of the crime rates in france , the dataset is not large enough ) the figure @xmath229(a ) . c + [ fig : ait ] this figure shows that in the case of the crime rate in the us , the @xmath17 converge to a stationary value , independent of the time interval , provided it is large enough . our method will then lead to reliable results constant in time . we also tested our method on the financial time series given by the 500 most important stocks in the us economy [ 1 ] , and which composition leads to the @xmath230 index . here the ` local ' units are the individual stocks ( @xmath231 ) , and the ( naive ) average - analogue to a national average - is precisely the @xmath232 index time serie . we study the time series for these stocks on the @xmath233 days of the period @xmath234 and we compute the global pattern @xmath16 , the coefficients @xmath17 , and the parameters @xmath131 ( defined in the text ) computed for the time window @xmath235 $ ] for @xmath31 varying from @xmath236 to @xmath237 . these quantities @xmath131 measure quantitatively the importance of local versus external fluctuations for the stock @xmath5 . the results for the @xmath131 s are shown in figure @xmath138(b ) and display large variations , particularly when we approach october , 2008 , a period of financial crisis . it is therefore not completely surprising that the @xmath131 ( and the @xmath17 s ) in this case fluctuate a lot . in some sense , we can conclude that the @xmath17 s correspond to an average susceptibility to the global trend , are not invariable quantities and can vary for different periods . we thus see on this example , that it is important to check the stability of the coefficients @xmath17 which is an crucial assumption in our method . the variations of these coefficients is however interesting and further studies are needed in order to understand these variations . [ 1 ] historical data for @xmath238 stocks http://biz.swcp.com/stocks/ for the obesity rate series , we compare the variances of the internal ( @xmath9 ) and the external ( @xmath239 ) contributions . we observe on the figure @xmath240 that the variance of the external contribution became dominant after the year @xmath241 .

a single social phenomenon ( such as crime , unemployment or birth rate ) can be observed through temporal series corresponding to units at different levels ( cities , regions , countries ... ) . units at a given local level may follow a collective trend imposed by external conditions , but also may display fluctuations of purely local origin . the local behavior is usually computed as the difference between the local data and a global average ( e.g. a national average ) , a view point which can be very misleading . we propose here a method for separating the local dynamics from the global trend in a collection of correlated time series . we take an independent component analysis approach in which we do not assume a small unbiased local contribution in contrast with previously proposed methods . we first test our method on synthetic series generated by correlated random walkers . we then consider crime rate series ( in the us and france ) and the evolution of obesity rate in the us , which are two important examples of societal measures . for crime rates , the separation between global and local policies is a major subject of debate . for the us , we observe large fluctuations in the transition period of mid-@xmath0 s during which crime rates increased significantly , whereas since the @xmath1 s , the state crime rates are governed by external factors and the importance of local specificities being decreasing . in the case of obesity , our method shows that external factors dominate the evolution of obesity since 2000 , and that different states can have different dynamical behavior even if their obesity prevalence is similar . : physical sciences ( applied mathematics , physics ) , social sciences . : time series analysis , independent component analysis , financial time series , crime rates , obesity .

from the early days of qcd , light meson spectroscopy played an important role in development of the theory and in understanding of its low energy structure . the flavour symmetry of qcd originates in part from the observed @xmath5 multiplet structure of the light pseudoscalar and vector mesons . more recently , scalar and tensor spectra have provided evidence for a possible over - population of the @xmath6 spectrum and for the existence of gluon - rich states . the possible existence of gluonic excitations is one of the most intriguing features of the meson spectroscopy . there is tantalizing evidence of exotic , hybrid mesons in the spectrum around @xmath7 [ 5 - 7 ] , and future experiments proposed for jlab and gsi in the light and charm sector respectively will map out the exotic spectrum . the glueball signature comes primarily from the analysis of the crystal barrel @xmath8 and wa102 central production data @xcite . detailed mapping of various scalar meson decay channels led to the identification of @xmath9 , @xmath10 and @xmath11 states , all expected to contain significant gluon components . while the genuine qcd resonances in the scalar channel are not expected to occur below 1 gev , the low energy region can be studied using standard , low energy expansion techniques . these include the effective range expansion , @xmath12 and other methods based on analyticity coupled with truncation of the unitarity condition . the parameters of the soft meson - meson interactions _ e.g. _ subtraction constants , form factors , coupling strengths , _ etc . _ effectively correspond to local potentials smeared over distance scales of the orders of @xmath13 , _ i.e. _ close to the pion root mean square radii . the strength of the interaction can be constrained from chiral symmetry or simply fitted to the data [ 8 - 11 ] . as a result one obtains a very good description of the spectrum including the resonance region . furthermore , the behavior of the scattering amplitude in the complex energy plane enables one to establish the existence of dynamical resonances such as the isoscalar @xmath14 and @xmath15 and the isovector @xmath16 mesons . the last two are particularly interesting as they are very sensitive to the details of the meson - meson interactions . this is due to the proximity of the @xmath17 threshold . in particular the detailed structure of the @xmath15 _ e.g. _ whether it is a genuine bound state or a virtual bound sate , strongly depends on the threshold @xmath17 interaction parameters . since the @xmath17 channel also influences the higher mass region a proper description of its dynamics is crucial for a partial wave analysis and the identification of exotic , hybrid or glueball signals . furthermore the @xmath17 system is relevant for testing the origins of cp violation and possible signals of cpt violation via hadronic interferometry @xcite . the data on the near threshold @xmath17 production are very scarce and come mainly from high - energy peripheral production @xcite . in the medium - energy region , @xmath18 , it is advantageous to study the @xmath17 system in photoproduction . a real photon couples strongly to vector mesons and near the @xmath17 threshold @xmath19 photoproduction dominates the @xmath17 spectrum . in this energy range the @xmath19 photoproduction cross section is large , @xmath20 , and has an energy dependence characteristic for diffraction . the @xmath21-channel baryon resonance production cross section is highly suppressed at these energies , and the @xmath22-channel meson exchange ( @xmath23 ) is marginally relevant only at @xmath24 very close to threshold @xcite . finally , the @xmath3-wave @xmath17 state , dominated in this mass range by the @xmath15 in the isoscalar and by the @xmath16 in the isovector channel , can be accessed via interference with the @xmath19 meson in the @xmath4-wave . this @xmath25 interference was first explored in experiments at desy @xcite and daresbury @xcite , and could be further studied with high statistics at an energy upgraded jlab . in the analysis of the data from the desy and daresbury experiments , however , the rich spectrum of the @xmath3-wave @xmath17 has not been fully explored . in both cases the @xmath3-wave was parameterized as either a simple breit - wigner resonance or a nonresonant background . the two analyses yielded results for the @xmath3-wave production cross section different from each other by more than one order of magnitude . finally , apart of the @xmath0 mass spectrum , only one moment @xmath26 , describing the @xmath27 angular distribution , has been analyzed and very simplified assumptions about the nucleon spin dependence have been done . in this paper we use the results of a recent calculation of the coupled channel scalar - isoscalar and scalar - isovector spectra together with a diffractive model for @xmath4-wave production to interpret the existing @xmath25 interference data and estimate the @xmath3-wave production cross section with a better accuracy than in the phenomenological analyses already published in [ 16 - 18 ] . here we shall discuss a complete set of six moments @xmath28 of spherical harmonics including @xmath29 and @xmath30 . in the following section we shall present the theoretical foundation of the @xmath3- and @xmath4-wave photoproduction and discuss the main features of the existing data . in sect . 3 we present results of the numerical analysis and fits to the data . conclusions and outlook are given in sect . in this paper we consider unpolarized @xmath17 photoproduction reaction @xmath33 for incident photon laboratory energies @xmath24 of the order of a few gev . this is an optimum energy range for the @xmath3-wave @xmath17 production . in this energy range the process is dominated by the pomeron exchange , leading to @xmath19-meson production which becomes even more important as the photon energy increases . however , @xmath22-channel processes expected to be responsible for the production decrease rapidly with energy . the experimental evidence for the @xmath25 interference in the @xmath24 range between 2.8 and 6.7 gev was presented in [ 16 - 18 ] . the values of the @xmath3-wave photoproduction cross sections found in these experimental analyses varied between 2.7 and 96 nb . in both cases , in the effective mass range @xmath34 , in the rest frame of the @xmath17 system an asymmetry in the kaon polar angle distribution was observed . this can only be the case if there are odd powers of @xmath35 in the angular distribution . for low partial waves this implies presence of both @xmath3- and @xmath4- waves . this feature of the angular distribution is independent of the magnetic - quantum number _ i.e. _ choice of the quantization axis . due to the empirical @xmath21-channel helicity conservation , diffractive production is most naturally analyzed in the @xmath36 helicity frame , which in our case corresponds to the rest frame of the @xmath17 pair with the @xmath37-axis anti - parallel to the direction of the recoiling nucleon . in another reference frame , called the @xmath22-channel frame or the gottfried - jackson frame , the @xmath37-axis is chosen along the direction of the photon beam with the @xmath17 pair at rest @xcite . in both cases the @xmath38-axis is perpendicular to the production plane . as discussed in @xcite , in this energy range , the @xmath3-wave production is expected to be dominated by vector @xmath39 , @xmath40 and pseudoscalar @xmath23 , @xmath41 @xmath22-channel exchanges . as a function of the momentum transfer squared @xmath22 , the kaon pair invariant mass @xmath42 and the @xmath27 decay angles @xmath43 in the @xmath21-channel frame the photoproduction amplitude restricted to @xmath3- and @xmath4-wave s can be written as @xmath44 here @xmath45 , @xmath46 and @xmath47 denote the photon , target proton and recoil proton helicities , respectively , and @xmath48 are the spherical harmonics . the corresponding four - momenta will be denoted by @xmath49 , @xmath50 , @xmath51 , and @xmath52 and @xmath53 will be used for the @xmath27 and the @xmath54 momenta , respectively . the unpolarized differential cross section is given by @xmath55 is the proton mass , @xmath56 is the kaon mass and @xmath57 is the kaon momentum in the rest system of the @xmath17 pair . the @xmath3-wave amplitude @xmath58 and the @xmath4-wave amplitude @xmath59 have been described in @xcite , respectively . here we will briefly summarize the basic properties of these amplitudes . the @xmath3-wave @xmath17 production is parameterized as a double exchange . in the upper meson vertex we use a simple meson exchange and allow for an interaction of the two produced mesons in the final state . the dominant exchanges for the @xmath3-wave @xmath17 production are shown in . at the nucleon vertex we use either normal or regge propagators of the exchanged vector mesons . the normal propagator of the vector meson of the mass @xmath60 is equal to @xmath61 and the regge type propagator reads @xmath62~\gamma(1 - \alpha(t))~ ( \alpha's)^{\alpha(t)}/(2s^{\alpha_0}),\ ] ] where we have the vector meson trajectory @xmath63 , @xmath64 and @xmath65 gev@xmath66 . 1 : the amplitude for the @xmath3-wave @xmath17 production . the @xmath67 in the born amplitude @xmath68 stands for @xmath23 , @xmath41 , @xmath39 and @xmath40 mesons , the @xmath69 stands for @xmath39 or @xmath40 mesons and @xmath70 denotes either a @xmath71 or @xmath17 pair . the oval represents the final state rescattering amplitude . the final state interactions include the @xmath71 and @xmath17 channels . both interactions can have a resonant character . this is important since the in the mass region of interest , @xmath72 , is dominated by the @xmath15 and @xmath16 resonances . we should notice that the @xmath0 system is an equal mixture of the isospin 0 and isospin 1 states . the isoscalar @xmath15 resonance has a main branching to @xmath71 and in addition an important coupling to @xmath17 . the isovector @xmath16 resonance lies also very close to the @xmath17 threshold so one should take it into account in the calculations of the final state interactions . the @xmath3-wave @xmath0 photoproduction amplitude @xmath73 can therefore be written as @xmath74 and decomposed as @xmath75 $ ] . here @xmath76 is the photon helicity four - vector and @xmath77 is the @xmath3-wave projection of the appropriate current operator . as illustrated in fig 1 each @xmath3-wave amplitude is a sum of products of the born amplitudes @xmath78 describing the @xmath22-channel meson or regge exchanges and the final state interaction factors @xmath79 @xmath80 if we restrict ourselves to the on - shell part of the final state coupled channel interactions , represented by the 2@xmath812 @xmath3-matrix elements @xmath82 , then the factors @xmath83 can be written as @xmath84 where @xmath85 and @xmath86 is the effective mass of the @xmath70 pion or kaon pair and @xmath87 is the pion or kaon mass . the explicit forms of the @xmath3-wave born amplitudes @xmath88 are given in @xcite . the magnitude of the off - shell part of the final state interaction amplitude is much less certain than the on - shell part as has been shown in @xcite . thus we prefer here to use only the on - shell part of the amplitude and correct it later by a constant modification factor . in comparison with @xcite we have added in these amplitudes a @xmath22- dependent factor @xmath89 , where the parameter @xmath90 gev@xmath66 is responsible for a spatial dimension of the meson ( kaon or pion ) coupled to the photon . in @xcite the meson in the upper vertex in fig . 1 was treated as a pointlike particle . the isoscalar @xmath82 matrix elements can be expressed by the scalar- isoscalar phase shifts @xmath91 and inelasticity @xmath92 in two channels @xmath71 and @xmath17 : @xmath93 these elements have been computed in @xcite and here we use the solution a corresponding to the so - called `` down - flat '' data of the phase shift analysis @xcite . the i=1 @xmath17 interaction near 1 gev is very strongly influenced by the @xmath16 resonance which decays dominantly into the @xmath94 . a coupled channel model including the @xmath94 and @xmath17 states has been formulated in @xcite and recently compared to the existing data @xcite . as a result one obtains the @xmath17 scalar - isovector @xmath3-matrix elements @xmath95 which differ from the scalar - isoscalar @xmath3-matrix elements written in ( [ si0 ] ) . this new information has been incorporated into calculations of the final state @xmath0 interactions . the @xmath4-wave amplitudes corresponding to a given projection m of the @xmath0 angular momentum on the quantization axis can be handled in a similar way as the @xmath3-wave , _ i.e. _ we write @xmath96 with @xmath97 being the @xmath4-wave projection of the current . the @xmath4-wave current is described in detail in @xcite . in particular , it is saturated by the diffractive @xmath19-meson production , @xmath98 , \label{p}\ ] ] where @xmath99 are the dirac matrices , and @xmath100 and @xmath101 are @xmath19 mass and width , respectively . the lorentz - dirac structure of the current is motivated by the donnachie - landshoff model for the pomeron exchange which assumes vector coupling of the pomeron to hadrons @xcite . the two terms in ( [ p ] ) are needed to preserve electromagnetic current conservation , @xmath102 . one can show that @xmath103 is satisfied independently by each projection of the @xmath17 spin , @xmath104 . then in the phenomenological analysis of the data one can separately modify different @xmath105 components by multiplying them by constant factors . the function @xmath106 is a phenomenological function which will be suitably parameterized to reproduce , at fixed energy , the @xmath22-dependence of the observed its analytical form will be specified in the next section . we note , however , that this model for the @xmath4-wave does not lead to significant suppression of the single helicity flip amplitude . in particular in the high energy limit of @xmath107 and @xmath108 the helicity non - flip , single - flip and double - flip amplitudes with @xmath109 behave as @xmath110 respectively . the proportionality constant @xmath111 contains the breit - wigner propagator of ( [ p ] ) and is finite as @xmath112 . the diagonal elements of the matrix @xmath113 are finite and the off - diagonal ones , corresponding to nucleon helicity flip , are @xmath114 . thus the only suppression of the photon - meson coupling comes from the angular momentum conservation factor @xmath115 , but it is otherwise finite at high energy . the existing data on @xmath19 photoproduction suggest that the helicity flip amplitude is of the order of @xmath116 of the dominant , helicity non - flip one at photon energies below @xmath117 @xcite . thus , qualitatively the model has the correct features but the quantitative agreement may require adjusting the normalization and the phase of the helicity flip amplitudes . this can be done by multiplication of these amplitudes by a constant complex factor @xmath118 , where @xmath119 is a real positive number and @xmath120 is an additional phase of the @xmath4-wave amplitude with @xmath121 . we will not modify phases of the dominant helicity non - flip amplitudes , however we do change slightly their moduli as well as the moduli of small double helicity flip amplitudes to keep the total @xmath4-wave cross - section untouched . the phases of the small double helicity flip amplitudes are the same as the phases of the corresponding non - flip amplitudes . similarly , for the @xmath3-wave amplitudes we introduce a constant modification factor equal to @xmath122 . it is assumed that the parameters @xmath123 and @xmath124 do not depend on proton helicities @xmath46 nor @xmath47 . the experimental data on the angular distribution of the @xmath17 pair are given in terms of the moments @xmath125 of the angular distribution evaluated in the @xmath21-channel helicity frame , @xmath126 where @xmath127 takes into account the photon flux and the 1/4 factor standing before the sum in ( [ ds / dtdmdo ] ) . the lowest moment is normalized to the @xmath0 mass distribution integrated over the momentum transfer squared range limited by @xmath128 and @xmath129 : @xmath130 in terms of the @xmath3- and @xmath4-partial waves , the non - vanishing moments are @xmath131 here @xmath3 , @xmath4 stand for @xmath58 and @xmath59 amplitudes , respectively , and summation over photon and nucleon spin indices is implicit , _ e.g. _ , @xmath132 the dominant @xmath133 and @xmath134 waves originating from @xmath135 and helicity non - flip production will manifest themselves in a large , positive @xmath136 and a large negative @xmath137 near @xmath138 the mass of the @xmath19 resonance . this is indeed the dominant feature of the data as shown in figs . 3 and 4 . since @xmath139 the @xmath3-wave is not expected to be significant in the mass spectrum _ i.e. _ in the @xmath140 moment . it will primarily contribute to the @xmath35 asymmetry measured by the @xmath141 and the @xmath142 moments . for diffractive @xmath4-wave production with equal phases of all it is expected that @xmath143 since the latter describes the interference between the @xmath3- and the helicity flip @xmath144-wave which has a smaller magnitude than @xmath133 one . the data suggests , however , that the two moments are comparable with more structure actually seen in the @xmath141 moment . this can only be possible if we allow a well defined pattern of phases in the three waves . this justifies our choice of additional parameters as already described above . in the following section we discuss the results of fitting this model to the experimental data . in the analysis of the @xmath3- and @xmath4-wave production we compare the model described above to the @xmath22 , @xmath42 and angular distribution of the @xmath17 system at two photon energies , @xmath145 @xcite and @xmath146 @xcite . the two photon energies represent averages of the photon beam energies used in the experiments performed at daresbury and at desy , respectively . at first we discuss the momentum transfer dependence of the cross section integrated over the @xmath17 effective mass and the kaon emission angles . at 4 gev photon energy the comparison between the experimentally measured differential cross section @xmath148 and the model cross section integrated over the @xmath17 mass is shown in fig . 2 : differential cross section at @xmath149 gev . the solid line shows the model for the @xmath19 photoproduction , the dotted line is the @xmath4-wave contribution with m=0 multiplied by the branching ratio of the @xmath19 decay into the @xmath0 pair . the dashed line is the @xmath3-wave part of the @xmath0 cross section calculated for normal @xmath150 propagators , while the double dotted - dashed line corresponds to the regge propagators . model parameters are given in data are from fig . 6b of @xcite . for this comparison , the @xmath22-dependent normalization @xmath106 in the @xmath4-wave was chosen as @xmath151 where the normalization constant @xmath152 has been adjusted to reproduce the very forward value of @xmath153gev@xmath154 at small argument , @xmath155gev@xmath154 , and the remaining parameters have been chosen as @xmath156 gev@xmath154 and @xmath157 gev@xmath66 . the value of @xmath158 has been obtained from the experimental fits of barber _ et al . _ at low momentum transfers and in the energy range between 3.4 and 4.8 gev ( see fig . 7 of @xcite ) . we took @xmath159gev@xmath154 @xmath160gev@xmath161 $ ] and calculated its value at the minimum @xmath162 argument corresponding to the @xmath17 effective mass 1.042 gev , which was the upper limit studied in @xcite . then the constant @xmath152 equals to @xmath163 where br=0.486 is the branching ratio for the @xmath19 decay into @xmath0 used in @xcite and @xmath164 @xmath165 being the lower effective mass limit 0.997 gev and @xmath166 being the upper mass limit 1.042 gev . in the above equation the unnormalized double differential cross section @xmath167 corresponds to the @xmath4-wave amplitudes calculated at fixed @xmath168 . a variation of the @xmath168 with the effective mass @xmath0 has been neglected in this narrow band of @xmath42 . as expected , the @xmath22-distribution is dominated by the helicity non - flip ve . the @xmath144-wave and the @xmath3-wave are kinematically suppressed at low @xmath22 and are two orders of magnitude smaller than the dominant wave . after integration over @xmath22 in the range up to @xmath169gev@xmath154 the total @xmath19 photoproduction cross section equals 0.449 @xmath170 and is in very good agreement with the measured value of @xmath171 @xmath170 presented in table 1 of @xcite for the photon energy range between 2.8 and 4.8 gev . the @xmath0 part of the integrated @xmath4-wave cross section equals @xmath172 @xmath170 . in the model fits to data , the decomposition of the total @xmath4-wave cross section in its parts corresponding to @xmath173 , @xmath121 and @xmath174 components depends to some extent on the contribution of the @xmath3-wave cross section and in particular on the choice of the type of propagators included in the @xmath3-wave amplitude . for normal @xmath150 propagators the integrated cross sections for the @xmath144-wave and the @xmath3-wave are equal to @xmath175 nb and @xmath176 nb , respectively . the corresponding numbers for the regge propagators are @xmath177 nb and the errors of the cross sections have been evaluated using the limits of the model parameters obtained from the fitting program minuit @xcite . these numbers are smaller than our early estimates of the @xmath3-wave integrated cross sections written in @xcite . one reason of this change is related to a presence of the @xmath3-wave form factor @xmath178 introduced in the previous section and the second one comes from the diminution of the @xmath3-wave modulus obtained in the fitting procedure which will be explained later . the desy data @xcite have been taken at much lower momentum transfers than the daresbury data @xcite . the most precise data of behrend et al . lie within the range @xmath180 gev@xmath154 . the average energy corresponds to the photon energy range between 4.6 and 6.7 gev . the @xmath19 production differential cross section @xmath148 can be fitted as , where @xmath181gev@xmath154 and @xmath182 gev@xmath66 represents the average slope of @xmath148 for the four upper energy bins given in table 3 of @xcite . taking the appropriate @xmath168 value equal to @xmath183gev@xmath154 we calculate @xmath184gev@xmath154 . then we define a simple form of the @xmath22-dependent normalization factor at @xmath185 gev as @xmath186 and use once again ( [ d1 ] ) and ( [ int ] ) to find a new constant @xmath187 taking into account that the @xmath19 branching ratio used in @xcite was br=1/2.14 , and @xmath188 gev . the @xmath0 photoproduction cross section at @xmath189 gev integrated over @xmath190 up to 0.2 gev@xmath154 is 120.5 nb . its @xmath144- and @xmath3-wave parts are @xmath191 nb and @xmath192 nb for normal propagators and @xmath193 nb and @xmath194 nb for regge propagators , respectively . let us notice that the @xmath144- and @xmath3-wave cross sections are comparable and do not vary too much with energy bearing in mind rather large errors . the @xmath3-wave cross section at 5.65 gev is comparable in its magnitude with the estimate of the upper limit @xmath195 nb quoted in @xcite and @xcite . we finally note that the total @xmath19 photoproduction cross section of @xmath196 @xmath170 given in @xcite corresponds to @xmath197 integrated over the range 1.0 gev to 1.024 gev . when integrated in this mass range our model gives 0.23 @xmath198 in good agreement with the measurement . next let us describe calculations of the @xmath0 differential cross section integrated over a certain range of the momentum transfer at fixed @xmath42 mass . we first discuss the results corresponding to @xmath145 where the upper limit of @xmath200 was 1.5 gev@xmath154 . .fitted values of model parameters for @xmath201 gev . units of @xmath202 and @xmath203 are @xmath204gev and @xmath204gev@xmath154 , respectively . [ cols="^,^,^",options="header " , ] we have performed calculations of the @xmath0 mass spectrum and moments at @xmath205=8 gev which will be a typical energy of the planned upgrade of the cebaf accelerator operating at the thomas jefferson laboratory . fig . 5 : @xmath0 mass spectrum and moments of angular distribution in the helicity frame for incident photon energy 8 gev . solid lines are results of model calculations for the normal @xmath150 propagators while the dashed lines correspond to the regge propagators in the @xmath206 wave amplitudes . the results presented in fig . 5 can be directly compared with fig . 3 corresponding to the much lower energy of 4 gev . the calculations have been performed for an ideal case in which there is no background and no phenomenological adjustment of the moduli and phases of the @xmath206 and @xmath207 waves i.e. @xmath208 , @xmath209 . we have assumed that the @xmath0 mass resolution is equal to 5 mev . the parameters of the @xmath210 meson , like the mass , width and the @xmath0 decay fraction , have been taken as 1019.456 mev , 4.26 mev and 0.492 , respectively . the differential cross section at low @xmath22 was parameterized as with @xmath211gev@xmath154 and @xmath212 gev@xmath66 . one can notice that the application of regge propagators in the @xmath206 wave leads to smaller values of the @xmath206 wave cross section and to smaller amplitudes of the moments @xmath213 and @xmath214 sensitive to the interference between the @xmath206 and @xmath215 waves . qualitatively we do not observe important differences in a behaviour of the unnormalized moments between the photon energy of 5.65 and 8 gev ( let us recall here that fig.4 shows the ratios of @xmath216 , not @xmath28 ) . the @xmath206 wave total cross sections integrated over the @xmath217 range between 1.01 and 1.03 gev and in the @xmath218 range from 0.0054 up to 0.2 gev@xmath219 are equal to 6.9 nb and 3.0 nb for the normal and regge propagators , respectively . the @xmath215 wave cross section integrated in the same ranges equals 141 nb while the corresponding @xmath207 cross section is equal to 6 nb . in this paper we presented results of a theoretical analysis of data on photoproduction of the @xmath0 pairs in the laboratory photon energy range @xmath24 between 2.8 and 6.7 gev . in particular we mapped out the interference pattern between the @xmath3- and the @xmath4-wave due to the presence of the scalar @xmath15 , @xmath16 and the vector @xmath210 resonances . in the analyses the @xmath3-wave was described by a model based on the dominance of the @xmath22-channel exchange process with the two - meson spectrum described in terms of the coupled channel meson - meson scattering @xmath3- matrix . the @xmath4-wave was described in terms of diffractive production of the @xmath19 resonance decaying into the @xmath0 system . the @xmath3-wave and the @xmath4-wave contain 4 and 12 independent amplitudes respectively and we included them all while previous analyses made a severe truncation to a single amplitude in each wave . we have shown that amplitudes omitted in the analyses of @xcite corresponding to the proton helicity non - flip are large and can not be ignored . in the previous analyses only the @xmath220 moment was taken into account and this led to a large variation in the estimate of the @xmath3-wave photoproduction cross section , between @xmath221 at @xmath222 @xcite and @xmath223 at @xmath224 @xcite . in this paper we have considered all six moments appearing in the model with @xmath3- and @xmath4-waves which enabled us to isolate the production cross section from that of an incoherent background . this led to a significant , order of magnitude , reduction in the uncertainty in the @xmath3-wave production and reduced the value of cross section at @xmath225 from @xmath226 to approximately @xmath227 , thus eliminating the discrepancy between the two measurements ( also suggested in @xcite ) . thus we have found that the @xmath3-wave photoproduction cross section integrated over the @xmath19 resonance region is between 4 and 7 nb for the two photon energies 4 and 5.65 gev . the cross section of the photon helicity - flip @xmath144-wave , which interferes with the @xmath3-wave is found to be somewhat larger than the @xmath3-wave cross section , in the range between 5 and 14 nb . new more precise measurements of the @xmath17 photoproduction with a _ simultaneous _ determination of the @xmath71 cross - section in the @xmath71 effective mass region near 1@xmath228 could provide new insight into the still controversial nature of the scalar mesons @xmath15 and @xmath16 . + * acknowledgments * we thank chueng - ryong ji and robert kamiski for their collaboration on the @xmath3-wave photoproduction and for an exchange of ideas on the @xmath4-wave amplitudes in the early stage of this work . one of us ( l. l. ) would like to thank maria raska and jacek turnau for enlightening discussions on the analysis of experimental data . this work was supported in part ( a. p. s. ) by the us department of energy grant under contract de - fg0287er40365 .

results of a new analysis of the @xmath0 photoproduction at two photon energies @xmath1 and @xmath2 with a particular emphasis on the @xmath3-wave production are presented . we show that the proper treatment of all the helicity components of the @xmath3- and @xmath4-waves enables one to eliminate the reported discrepancies in the extraction of the @xmath3-wave photoproduction cross section from experimental data . = 19.2 cm

the distribution amplitude ( da ) is the key factor for exclusive processes . usually , the das of the light pseudoscalar mesons can be expanded in gegenbauer polynomials , and their corresponding gegenbauer moments have been studied by various groups , cf . refs.@xcite . however , there is no definite conclusion on whether the shape of the da is in an asymptotic form @xcite or in a more broad form @xcite . the light pseudoscalar meson - photon transition form factor @xmath17 that describes the effect of the strong interaction on @xmath0 transition ( @xmath1 stands for @xmath2 , @xmath3 and @xmath4 ) provides a good platform for studying the leading - twist da , since it contains only one bound state and the power suppressed light meson s higher helicity and higher twist structures usually give negligible contributions . one can extract useful information on the leading - twist da by comparing the theoretical with the measured data on @xmath17 . based on cello , cleo and the babar data on @xmath18 @xcite , many people have discussed the properties of pion da @xcite . a cz - like da or even a flat da @xcite can explain the large @xmath9 behavior shown by the babar data @xcite , however the theoretical estimation on the form factor with a cz - like or flat da shall always be lower than the experimental results in small @xmath9 region @xcite . while by taking the non - valence quark parts into consideration , it is found that by setting the second pion moment @xmath19 around @xmath20 , one can explain the behavior in small @xmath9 region well , however a somewhat large discrepancy emerges in high @xmath9 region if the babar data is confirmed @xcite . a reasonable theoretical estimation on the form factor should explain the measured form factor s behaviors in both the lower and higher @xmath9-regions consistently . and one should find a way to compare the experiment results on the form factor to determine which da shape is more suitable . more over , the experimental data on @xmath21 or @xmath4 transition @xcite , especially , the new babar results within the region of @xmath22 $ ] gev@xmath23 @xcite , can provide further constraints on the pseudoscalar meson s da @xcite . then , we shall have three pseudoscalar meson - photon transition form factors @xmath17 to constrain the light meson s wavefunction ( wf ) / da parameters . because of @xmath3 and @xmath4 mixing , their condition is somewhat more difficult than the pionic case . two mixing schemes are adopted in the literature , even though their mixing parameters can be related through the correlation given by refs.@xcite , there are differences in dealing with certain processes . one mixing scheme is based on the flavor singlet @xmath24 and octet @xmath25 , under which , one usually introduces two mixing angles @xmath26 and @xmath27 and adopts the same da for @xmath24 and @xmath25 @xcite . in the present paper , we adopt the simpler quark - flavor mixing scheme @xcite , which is based on the quark - flavor basis @xmath28 and @xmath29 and only one mixing angle @xmath13 is introduced . since @xmath28 and @xmath29 have similar structure as that of pion , it is natural to adopt the same wf model for @xmath2 , @xmath28 and @xmath29 . for the purpose , we adopt the wf model raised by ref.@xcite to do our discussion , since by setting the parameter @xmath12 properly , we can obtain different da behavior from as - like to cz - like naturally , and then to determine which one is more suitable for simultaneously explaining the experimental data of these form factors . the paper are organized as follows . in sec.ii , we outline our calculation techniques for the transition form factors @xmath17 , where the mixing scheme for @xmath3 and @xmath4 , an uniform wf model for the mesons , and the analytic formulas for deriving @xmath17 are presented . in sec.iii , we present the numerical results , and discuss the uncertainty sources . the final section is reserved for a summary . in this section , we present necessary formulas for the transition form factors @xmath17 . first , we define the physical meson states @xmath3 and @xmath4 under the quark - flavor basis , then we give the uniform wf model for the mentioned pseudoscalar mesons , and finally , we present the form factor with the valence quark contribution calculated up to next - to - leading order ( nlo ) together with an estimation of the non - valence quark state s contributions . the physical meson states @xmath3 and @xmath4 are related to the orthogonal states @xmath28 and @xmath29 through an orthogonal transformation @xmath30 where @xmath13 is the mixing angle . here , we adopt a single mixing angle scheme that attributes @xmath31 breaking to the okubo - zweig - iizuka violating contribution @xcite . in the quark - flavor basis , the two orthogonal states @xmath32 and @xmath33 are defined in a fock state description , @xmath34 and @xmath35 , where @xmath36 ( @xmath37 , @xmath38 ) denote the light - cone wfs of the corresponding parton states . under such scheme , the decay constants in the quark - flavor basis simply follow the pattern of state mixing , i.e. @xmath39 , \label{decayconstant}\end{aligned}\ ] ] where @xmath40 and the factorization scale @xmath41 . one can obtain the correlation between @xmath42 and @xmath13 from the two - photon decay of @xmath3 and @xmath4 , i.e. @xmath43 , which shows @xcite @xmath44^{-1}\end{aligned}\ ] ] and @xmath45^{-1 } , \label{decayfs}\end{aligned}\ ] ] where @xmath46 , @xmath47 and @xmath48 . since the power suppressed higher twist and higher helicity components give negligible contributions to the two - photon decay of @xmath3 and @xmath4 , the correlations ( [ decayfq],[decayfs ] ) shall provide strong constraint on @xmath49 , @xmath50 and @xmath13 . as for the light pseudoscalar meson ( @xmath1 ) , its light - cone wf can be written as @xmath51 where @xmath52 stands for the light quark @xmath53 or @xmath38 , @xmath54 and @xmath55 are helicities of the two constituent quarks . @xmath56 stands for the spin - space wf coming from the wigner - melosh rotation . @xmath57 stands for the spatial wf , which can be factorized as @xcite @xmath58.\ ] ] the @xmath59-dependent @xmath60 can be expanded in gegenbauer polynomials , and by keeping its first two terms , we obtain @xmath61 .\ ] ] as for the parameters @xmath12 and @xmath62 , @xmath12 determines the broadness of the wf in the longitudinal direction , while @xmath62 determines the wf s transverse behavior . the da @xmath63 can be obtained from @xmath64 by integrating over the transverse momentum , @xmath65 , and we obtain @xmath66- \mathrm{erf}\left[\sqrt{\frac{m^2}{8\beta^2 x(1-x)}}\right ] \right ) , \label{ourphi}\end{aligned}\ ] ] where @xmath67 . @xmath68 can be expanded in conventional gegenbauer polynomials , whose moments @xmath69 ^ 2}$ ] . numerically , it is found that its second gegenbauer moment @xmath70 is close to @xmath12 , i.e. the da s behavior is dominated by @xmath12 . moreover , when @xmath71 , its da is asymptotic - like ; and when @xmath72 , its da is cz - like . this shows @xmath68 can mimic the da behavior from asymptotic - like to cz - like naturally . then , by comparing the estimations for @xmath73 $ ] with the experimental data on various processes , one may decide which is the right da behavior possessed by the light pseudoscalar mesons . here , we do not discuss the flat da , since it is hard to explain the meson - photon transition form factor s behavior around @xmath74 and shall meet even more serious end - point problem than the cz - like da at @xmath75 @xcite . the pseudoscalar meson - photon transition form factors can be divided into two parts @xmath76 where @xmath77 is the valence quark part , @xmath78 stands for the non - valence quark part that is related to the higher fock state of the pseudoscalar meson . under the light - cone perturbative qcd approach @xcite , and by keeping the @xmath79-corrections in both the hard - scattering amplitude and the wf , @xmath80 has been calculated up to nlo @xcite . for pseudoscalar meson - photon transition form factors up to nlo , we have @xcite , @xmath81 \cdot \nonumber\\ & & \psi_{p}(x , k_\perp^2 ) d k^2_\perp , \label{ffv}\end{aligned}\ ] ] where @xmath82=dxdx'\delta(1-x - x')$ ] , @xmath83 and @xmath84 . and @xmath85 relates to the electric charge of the constituent quarks , @xmath86 , @xmath87 and @xmath88 . as for @xmath89 , we adopt the model suggested by refs.@xcite , which is constructed based on the form factor s limiting behavior at both @xmath90 and @xmath91 , i.e. , @xmath92 where @xmath93 , @xmath94 with the first derivative of @xmath89 over @xmath9 in the limit @xmath90 takes the form @xmath95_{q^2\to 0}.\ ] ] the same phenomenological model for @xmath89 has also been adopted by ref.@xcite , where instead a fixed input parameter @xmath96 gev is introduced to replace the parameter @xmath97 . since the octet - singlet mixing scheme @xcite is adopted by ref.@xcite , it is reasonable to take the same @xmath98 for @xmath2 , @xmath25 and @xmath99 . under the present adopted quark - flavor mixing scheme , we shall numerically obtain @xmath100 gev for pion and @xmath28 , and @xmath101 gev for @xmath29 , where different @xmath97 is rightly caused by @xmath102-breaking effect . moreover , under the quark - flavor mixing scheme , @xmath103 and @xmath104 transition form factors are related with @xmath105 and @xmath106 through the following equations @xmath107 and @xmath108 two photon decay widths of @xmath3 and @xmath4 , and their masses can be found in pdg @xcite @xmath109 and @xmath110 mev . a weighted average of the experimental values shown in ref.@xcite , together with two experimental values @xmath111 @xcite and @xmath112 @xcite , yields @xmath113 . then , with the help of eqs.([decayfq],[decayfs ] ) , we obtain @xmath114 and @xmath115 which are consistent with the phenomenological " values , @xmath116 and @xmath117 @xcite . .typical wf parameters for @xmath118 gev and @xmath119 gev , where @xmath120 is adopted . [ cols="^,^,^,^,^",options="header " , ] [ tab1 ] as for pion wf , its parameters can be determined by its normalization condition and the constraint from @xmath121 . a detailed determination on pion wf parameters can be found in ref.@xcite , where @xmath122 for specified @xmath12 and quark mass is determined by @xmath123 because @xmath28 and @xmath29 have similar behaviors as that of @xmath2 , for clarity , we take @xmath124 and @xmath125 . under the condition of @xmath126 , @xmath127 and @xmath128 , typical parameters for the das of @xmath2 , @xmath28 and @xmath29 are presented in tab.([tab1 ] ) , where the mixing angle @xmath13 is fixed to be @xmath129 . ) for @xmath130 ( left ) and @xmath131 ( right ) with @xmath132 gev , where @xmath126 , @xmath127 and @xmath128 respectively . , title="fig:",scaledwidth=45.0% ] ) for @xmath130 ( left ) and @xmath131 ( right ) with @xmath132 gev , where @xmath126 , @xmath127 and @xmath128 respectively . , title="fig:",scaledwidth=45.0% ] it is noted that by varying @xmath12 within the region of @xmath133 $ ] , the das shall vary from asymptotic - like to cz - like form . to show this point more clearly , we draw @xmath130 and @xmath131 in fig.([phi ] ) , where @xmath126 , @xmath127 and @xmath128 respectively . numerically , one may observe that in the large @xmath9 region , the leading valence fock - state contribution dominates the form factor @xmath134 , while the non - valence quark part @xmath135 is power suppressed and is quite small , so it is usually neglected in the literature . however , @xmath135 can provide sizable contributions in the low and intermediate energy regions , so one should take it into consideration to make a more sound estimation in the whole energy regions . as shown above , the parameter @xmath12 in the unified wf model ( [ wave ] ) determines the da behavior of the light pseudoscalar mesons . then , by comparing with the experimental data on the pseudoscalar meson - photon transition form factors , it provides us an opportunity to discuss the da properties in a more consistent way . transition form factor @xmath15 with varying @xmath12 . the solid , the dotted and the dashed lines are for @xmath126 , @xmath127 and @xmath128 respectively . the experimental data are taken from refs.@xcite.,scaledwidth=50.0% ] ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 . the solid , the dotted and the dashed lines are for @xmath126 , @xmath127 and @xmath128 respectively . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 . the solid , the dotted and the dashed lines are for @xmath126 , @xmath127 and @xmath128 respectively . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] firstly , we present the meson - photon transition form factors within a wide region of @xmath73 $ ] so as to show which da behavior is more suitable to explain the data , especially the babar data @xcite . in doing the numerical calculation , we take all the other input parameters to be their center values , i.e. @xmath118 gev , @xmath119 gev and @xmath120 . figs.([pion ] , [ eta ] ) show the pseudoscalar - photon transition form factors @xmath15 , @xmath7 and @xmath8 , where @xmath126 , @xmath127 and @xmath128 respectively . these two figures show that with the increment of @xmath12 , all the three form factors decrease in lower @xmath9 region but increase in higher @xmath9 region . especially , the cz - like da ( @xmath136 ) leads to the smallest value in the lower @xmath9 region , while the as - like one ( @xmath126 ) leads to biggest one ; and in the higher @xmath9 region , the condition is vice versa . this causes the present puzzle for explaining the newly obtained babar data on @xmath16 form factor . the babar data shows that in the range of @xmath137 $ ] gev@xmath23 , the pion - photon transition form factor behaves as @xcite , @xmath138 , where @xmath139 and @xmath140 . a cz - like da with @xmath141 or even flat da can explain the data well for higher @xmath9 region , , i.e. @xmath142 @xcite , which is close to babar data . ] however it fails in lower @xmath9 region . the as - like da with @xmath143 provides a better understanding for lower @xmath9 region , however due to the fact that @xmath144 tends to be a constant value @xmath145 , it can not explain the large @xmath9 behavior . on the one hand , by increasing @xmath12 from @xmath146 to a larger value , the estimated large @xmath9-behavior of @xmath144 can be improved . on the other hand , the deviation of the lower @xmath9-behavior also increases with the increment of @xmath12 , so @xmath12 should not be too large . moreover , as shown by fig.([eta ] ) , @xmath103 form factor @xmath7 prefers a da with smaller @xmath12 , i.e. @xmath147 . for the @xmath104 form factor @xmath8 , all das shall lead to large @xmath9 behavior well above the experimental data . transition form factor @xmath15 with fixed @xmath148 . the red solid line , the blue dotted line and the black dashed line are for @xmath149 gev , @xmath127 gev and @xmath20 gev respectively . the experimental data are taken from refs.@xcite . , scaledwidth=45.0% ] ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with fixed @xmath148 and @xmath120 . the red solid line , the blue dotted line and the black dashed line are for [ @xmath149 gev and @xmath150 gev ] , [ @xmath118 gev and @xmath119 gev ] , and [ @xmath151 gev and @xmath152 gev ] respectively . the experimental data are taken from refs.@xcite , title="fig:",scaledwidth=45.0% ] ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with fixed @xmath148 and @xmath120 . the red solid line , the blue dotted line and the black dashed line are for [ @xmath149 gev and @xmath150 gev ] , [ @xmath118 gev and @xmath119 gev ] , and [ @xmath151 gev and @xmath152 gev ] respectively . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] secondly , we study the uncertainties of the transition form factors caused by @xmath10 and @xmath11 . for the purpose , we take @xmath153 gev and @xmath154 gev , and fix the parameter @xmath155 and @xmath120 . the @xmath16 form factor for @xmath153 gev is presented in fig.([pionmassuncern ] ) , and @xmath103 and @xmath104 form factors for [ @xmath149 gev and @xmath150 gev ] , [ @xmath118 gev and @xmath119 gev ] , and [ @xmath151 gev and @xmath152 gev ] are presented in fig.([etamassuncern ] ) . it can be found that the form factors change with the constituent quark mass similar to its change with @xmath12 , i.e. with the increment of constituent quark mass , all the three form factors increase in the lower @xmath9 region and decrease in the higher @xmath9 region . naively , one may expect to obtain a larger high @xmath9 behavior by setting a smaller @xmath10 and @xmath11 . especially , by setting the limiting values of @xmath156 and @xmath157 , which rightly corresponds to a flat @xmath158 as suggested by ref.@xcite , one can obtain the same logarithmic behavior for large @xmath9 region that is consistent with babar data , i.e. @xmath159 with @xmath160 . however , it is found that @xmath10 can not be too small , e.g. it should be larger than @xmath161 gev , otherwise the probability of leading valence quark state @xmath162 shall be larger that @xmath163 @xcite . ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with fixed @xmath148 and @xmath118 gev and @xmath119 gev . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with fixed @xmath148 and @xmath118 gev and @xmath119 gev . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] the conditions for @xmath103 and @xmath104 are somewhat more different . due to @xmath164 mixing , we need to consider these two form factors simultaneously . the curves for @xmath7 and @xmath8 for @xmath165 are presented in fig.([phiuncern ] ) . it is found that @xmath166 decreases with the increment of @xmath13 , while @xmath8 increases with the increment of @xmath13 . by shifting @xmath13 to a smaller value @xmath167 , one can explain these two form factors within @xmath168 gev@xmath23 consistently as has been pointed out by ref.@xcite . however such a shifting of @xmath13 can not explain the newly babar data on @xmath103 and @xmath104 for even larger @xmath169 gev@xmath23 . so it is hard to fit the gap between the theoretical estimation and the experimental data in the whole @xmath9 region by a simple variation of @xmath13 . experimentally , @xmath7 still increases with the increment of @xmath9 up to a large value even though its ascending trends is slower than the growth of @xmath15 , while @xmath8 tends to be a consistent for @xmath91 . so some other sources have to be introduced to explain both the @xmath103 and @xmath104 form factors in the whole energy region consistently . as shown by fig.([eta ] ) , @xmath170 can agree with the data with @xmath6 , so we hope the new sources shall have less effects to @xmath170 than that of @xmath171 . it has been suggested that a proper intrinsic charm component may have some help to explain the abnormally large production of @xmath4 @xcite . in the following subsection , we shall make a detailed discussion on the possible contributions from the intrinsic charm components . since the mixing between the @xmath174 state with @xmath175-@xmath176 basis is quite small @xcite , we can set @xmath177 where @xmath178 and @xmath179 corresponds to the contributions from the intrinsic charm component in @xmath3 and @xmath4 respectively . similarly , the wf of the intrinsic " charm component @xmath180 can be modeled as @xmath181,\end{aligned}\ ] ] where we adopt @xmath182 . the overall factor @xmath183 is determined by the wf normalization , in which their corresponding decay constants @xmath184 and @xmath185 are related by @xcite , @xmath186 $ ] . here to calculate @xmath178 and @xmath179 , the charm quark mass effect should be taken into consideration in the hard part of the amplitude , i.e. the higher helicity states that are proportional to the quark mass shall provide sizable contributions . after integrating over the azimuth angle , a direct calculation shows @xcite @xmath187 k_\perp d k_\perp , \ ] ] where @xmath188 and @xmath189 . ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with varying @xmath190 , where we take @xmath155 , @xmath118 gev , @xmath119 gev and @xmath191 gev . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with varying @xmath190 , where we take @xmath155 , @xmath118 gev , @xmath119 gev and @xmath191 gev . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] taking @xmath155 , @xmath118 gev , @xmath119 gev , @xmath191 gev and @xmath120 , we show how @xmath14 affects the form factors @xmath172 and @xmath173 . the results are presented in fig.([charma ] ) , where @xmath192 , @xmath193 mev , @xmath194 mev and @xmath195 mev respectively . these two form factors are slightly affected by the charm component in low @xmath9 region , while in high @xmath9 region , the form factors are quite sensitive to @xmath190 and they can be greatly suppressed by possible charm component . one may observe that the experimental data disfavors a larger portion of charm component as @xmath196 mev . and for a more larger @xmath197 , it shall have more obvious effects to @xmath171 than that of @xmath170 , which is what we wanted . ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with varying @xmath12 , where @xmath198 mev , @xmath118 gev , @xmath119 gev and @xmath191 gev . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] ( left ) and @xmath104 ( right ) transition form factors @xmath7 and @xmath8 with varying @xmath12 , where @xmath198 mev , @xmath118 gev , @xmath119 gev and @xmath191 gev . the experimental data are taken from refs.@xcite . , title="fig:",scaledwidth=45.0% ] more over , we present the results for @xmath103 and @xmath104 transition form factors with fixed @xmath198 mev in fig.([charmb ] ) , where @xmath199 , @xmath127 and @xmath128 respectively . it shows that with proper charm component @xmath5 mev and @xmath6 , the experimental data on @xmath7 and @xmath8 can be consistently explained . it is found that @xmath198 mev is consistent with ref.@xcite . because we still have @xmath200 mev , according to the mass - matrix - element shown by ref.@xcite , we still have @xmath201 up to high accuracy . and applying the parameters into the formulas presented in ref.@xcite , it can be found that the mixing between @xmath174 state with @xmath202 basis is still quite small , i.e. the mixing is around @xmath203 . then our present approximations ( [ qsc1],[qsc2 ] ) are still reasonable . light pseudoscalar meson - photon transition form factor provides a good platform to study the leading - twist da of the light pseudoscalar mesons since it contains only one bound state . in the present paper , we have analyzed three pseudoscalar meson - photon transition form factors consistently by using an uniform wf model suggested by ref.@xcite . by comparing the estimations with the experimental data on these form factors , it can provide strong constraints on the light pseudoscalar meson das . our results are listed in the following . \(1 ) according to eqs.([fft],[ffv],[ffnv ] ) , all pseudoscalar meson - photon transition form factors @xmath15 , @xmath204 and @xmath205 should have similar behaviors . since no rapid growth of @xmath7 and @xmath8 as that of @xmath15 has been found experimentally , these two form factors can be explained by setting @xmath6 together with small charm quark component @xmath206 mev . such a moderate da with @xmath6 for @xmath28 and @xmath29 , which corresponds to the second gegenbauer da moment around @xmath20 , may also be the pion da behavior . \(2 ) we have made a detailed discussion on the form factors uncertainties caused by the constituent quark masses @xmath10 and @xmath11 , the parameter @xmath12 , the mixing angle @xmath13 and the possible intrinsic charm components @xmath207 and @xmath14 . firstly , the parameter @xmath12 determines the main behavior of the form factors . by varying @xmath73 $ ] , one can conveniently obtain the results for @xmath16 transition form factor with those obtained in the literature , in which da behavior varies from as - like to cz - like accordingly . a cz - like da with @xmath141 can explain the data in high @xmath9 region , however it fails to explain the form factors lower @xmath9 behavior . while the as - like da with @xmath208 can give a better understanding for lower @xmath9 region , it can not explain the present babar data for large @xmath9 behavior . secondly , the parameters @xmath10 , @xmath11 and @xmath13 can improve the behavior slightly . with the increment of @xmath10 and @xmath11 , three form factors @xmath15 , @xmath7 and @xmath8 increase in lower @xmath9 region and decrease in higher @xmath9 region . @xmath166 decreases with the increment of @xmath13 , while @xmath8 increases with the increment of @xmath13 . thirdly , @xmath7 and @xmath8 are slightly affected by the charm component in low @xmath9 region , while in high @xmath9 region , the form factors are quite sensitive to @xmath190 and they can be greatly suppressed by possible charm component . \(3 ) it has been found that by adjusting these parameters within their reasonable regions , one can improve the estimations of the form factors to a certain degree but still can not solve the puzzle , especially to explain the behavior of @xmath16 form factor within the whole @xmath9 region consistently . due to the cancelation between @xmath209 and @xmath210 , it is reasonable that @xmath8 tends to be a constant as @xmath211 @xcite . however , it is hard to understand why @xmath7 and @xmath15 have such a quite different large @xmath9 behavior . especially it is hard to explain the rapid growth of @xmath15 , i.e. probably the logarithmic growth @xcite , reported by babar collaboration @xcite to be consistent with the previously obtained low @xmath9 behavior by cello and cleo collaborations @xcite . possible charm components @xmath207 and @xmath14 can shrink the gap between these two form factors to a certain degree , but it can not be the reason for such a large difference . we hope more experimental data on these form factors in large @xmath9 region can clarify the present situation . if the babar data is confirmed , then there may indeed indicate new physics in these form factors , since it is hard to be explained by the current adopted light - cone pqcd framework . * acknowledgments * : this work was supported in part by the fundamental research funds for the central universities under grant no.cdjzr101000616 and the program for new century excellent talents in university under grant no . ncet-10 - 0882 , and by natural science foundation of china under grant no.10975144 , no.10805082 and no.11075225 . bakulev , s.v . mikhailov and n.g . stefanis , phys.lett . b*578 * , 91(2004 ) ; a.p . bakulev , k. passek - kumericki , w. schroers and n.g . stefanis phys.rev . d*70 * , 033014(2004 ) ; erratum - ibid.d*70 * , 079906(2004 ) ; a.p . bakulev , s.v . mikhailov and n.g . stefanis , phys.rev . d*67 * , 074012(2003 ) ; a.p . bakulev , s.v . mikhailov and n.g . stefanis , phys.rev . d*73 * , 056002(2006 ) . brodsky , t. huang and g.p . lepage , in _ particles and fields-2 _ , proceedings of the banff summer institute , banff , alberta , 1981 , edited by a.z . capri and a.n . kamal ( plenum , new york , 1983 ) , p143 ; t. huang , _ in proceedings of xxth international conference on high energy physics _ , madison , wisconsin , 1980 , edited by l.durand and l.g . pondrom , aip conf.proc.no . 69(aip , new york , 1981 ) , p1000 .

the meson - photon transition form factors @xmath0 ( @xmath1 stands for @xmath2 , @xmath3 and @xmath4 ) provide strong constraints on the distribution amplitudes of the pseudoscalar mesons . in this paper , these transition form factors are calculated under the light - cone perturbative qcd approach , in which both the valence and non - valence quarks contributions have been taken into consideration . to be consistent , an unified wavefunction model is adopted to analyze these form factors . it is shown that with proper charm component @xmath5 mev and a moderate da with @xmath6 , the experimental data on @xmath7 and @xmath8 in whole @xmath9 region can be explained simultaneously . further more , a detailed discussion on the form factors uncertainties caused by the constituent quark masses @xmath10 and @xmath11 , the parameter @xmath12 , the mixing angle @xmath13 and @xmath14 are presented . it is found that by adjusting these parameters within their reasonable regions , one can improve the form factor to a certain degree but can not solve the puzzle for @xmath15 , especially to explain the behavior of @xmath16 form factor within the whole @xmath9 region consistently . we hope further experimental data on these form factors in the large @xmath9 region can clarify the present situation . pacs numbers : : 12.38.-t , 12.38.bx , 14.40.be

this paper considers the issue of fubini s theorem and non - linear polynomial changes of variables for integration over a two - dimensional local field ( that is , a complete discrete valuation field @xmath0 with residue field @xmath1 a usual local field ) . integration over such fields was first considered by fesenko @xcite , then by kim and lee @xcite . hrushovski and kazhdan @xcite @xcite then developed a very general integration theory on valued fields using model theory ; their methods are valid for residue characteristic zero ( and sufficiently large positive characteristic ) . the author @xcite has reformulated fesenko s work to provide a new perspective which appears to be useful in treating problems such as fubini s theorem . a satisfactory theory of translation - invariant integration on algebraic groups over two dimensional local fields should lead to significant advances in the representation theory of such groups . kim and lee consider this problem in @xcite and construct a translation - invariant measure on @xmath2 . the methods of hrushovski and kazhdan yield all main results under the restriction on the characteristic . the author extended @xcite his methods of @xcite to obtain a translation invariant integral on @xmath2 which lifts in a natural sense the usual haar integral on @xmath3 . a relation between this theory and quantum physics is evidenced by the two - dimensional local field @xmath4 . subspaces of this space such as @xmath5 $ ] may be identified with subspaces of the space of continuous paths in the plane . the feynman integral is not understood rigorously ( see @xcite for discussion of the problems ) and measure theory on these path spaces may help to explain it . further evidence of the relations between quantum field theory and the measure on these fields may be found in sections 16 , 18 of @xcite and example 5.6 of @xcite . to extend the approach in @xcite from @xmath6 to an arbitrary algebraic group it is necessary to have a theory of integration on finite dimensional vector spaces over @xmath0 which behaves well under certain _ non - linear _ changes of variable ( for the @xmath6 theory , linear changes of variable suffice ) . moreover , for use in applications , it is essential that fubini s theorem concerning repeated integrals is valid . this paper considers the problem of establishing whether the equality @xmath7 holds for appropriate functions @xmath8 and polynomials @xmath9 . moreover , the methods used appear to be suitable for changes of variables much more general than @xmath10 . in the first section of the paper we summarise the integration theory of @xcite . given a haar integrable function @xmath11 , let @xmath12 be the lift of @xmath13 to @xmath0 which vanishes outside @xmath14 and satisfies @xmath15 for @xmath16 ; functions expressible as finite sums of functions of the form @xmath17 for @xmath18 , @xmath19 , @xmath13 haar integrable on @xmath1 , are said to be integrable on @xmath0 . a translation invariant @xmath20-valued integral @xmath21 is defined for such functions . in the second section we discuss the extension of this integration theory to @xmath22 . given a haar integrable function @xmath23 , let @xmath12 be the lift to @xmath22 which vanishes outside @xmath24 . taking translations and scaling of such function by @xmath22 and @xmath25 obtains a space of functions similar to the analogous space for @xmath0 . however , is appears that there are many more functions @xmath26 for which @xmath27 that is , for which both repeated integrals make sense according to the integration theory of @xmath0 and the two repeated integrals have the same value . the third section describes the action of polynomials on @xmath0 . given a polynomial @xmath28 $ ] , and a translated fractional ideal @xmath29 , we show how to write @xmath30 as a finite disjoint union of translated fractional ideals ; here @xmath31 is a local parameter of @xmath0 as a discrete valuation field . if @xmath32 is one of these translated fractional ideals , it is also important to understand the behaviour of the function @xmath33 these results ( indeed , the whole paper ) are closely related to ramification theory of two - dimensional local fields through a. abbes and t. saito s theory of ramification for complete discrete valuation fields with imperfect residue field ; a summary of their theory is given towards the end of section three , and further relations explained throughout the remainder of the paper . the impetus of this paper is conjecture [ conjecture ] , which we rapidly reduce to the following : if @xmath13 is a schwartz - bruhat function on @xmath34 , @xmath12 is defined as above , and @xmath35 $ ] is a polynomial , then surely @xmath36 in section five the conjecture is shown to be true if @xmath9 is linear or if all coefficients of @xmath9 belong to @xmath14 . the technically difficult case of when @xmath9 contains coefficients not in @xmath14 is taken up in section six . introduce a polynomial @xmath37 $ ] and integer @xmath38 by the three conditions @xmath39 , @xmath40 $ ] , and @xmath41 $ ] . we give explicit expressions for the integral of @xmath42 in terms of the decomposition of sets of the form @xmath43 ; the conjecture easily follows if @xmath44 so long as @xmath45 , the image of @xmath46 in @xmath47 $ ] , is not a purely inseparable polynomial . when @xmath48 calculations become difficult , and the function @xmath49 can fail to be integrable , meaning that the conjecture fails ; however , we present examples suggesting that the space of integrable functions could be extended so as to remedy this deficit . we then consider the possibility that @xmath1 has positive characteristic and @xmath45 is purely inseparable . when @xmath44 it is shown that @xmath50 but @xmath51 so if @xmath13 has non - zero haar integral over @xmath34 then the conjecture drastically fails . this fascinating result provides an explicit example to show that the work of hrushovski and kazhdan really can fail in positive characteristic , and we discuss its relationship with ramification theory . in the final section we summarise the results obtained and discuss future work . the bulk of this paper was written during the first half of my ph.d . studies under i. fesenko , and i am grateful to him for many interesting discussions surrounding adlic analysis . later , g. yamashita kindly read the first version of this paper in detail and introduced me to the ramification theory of a. abbes and t. saito , which i began to study while travelling abroad in the final year of my phd . these visits abroad would not have been possible without the generosity of the cecil king foundation and london mathematical society , who funded me through the cecil king travel scholarship . chapter @xmath52 of my thesis @xcite consists mainly of the material of this paper , but with additional comments ; the reader is recommended to consult that source in conjunction with the current paper . @xmath0 is a complete discrete valuation field whose residue field @xmath53 is a local field ( @xmath54 , @xmath55 , or non - archimedean ) . we fix a prime @xmath31 of @xmath0 and denote its ring of integers by @xmath14 . the residue map @xmath56 is denoted @xmath57 ; the discrete valuation is denoted @xmath58 . we fix a haar measure on @xmath59 . in @xcite a theory of integration on @xmath0 taking values in the field of rational functions @xmath20 is developed . we repeat here the definitions and main results : let @xmath60 denote the space of complex - valued haar integrable functions on @xmath53 . let @xmath61 denote the smallest @xmath20 space of all @xmath20-valued functions on @xmath0 with the following properties : 1 . for all @xmath62 , @xmath61 contains @xmath63 2 . if @xmath64 and @xmath19 then @xmath61 contains @xmath65 ; 3 . if @xmath64 and @xmath18 then @xmath61 contains @xmath66 . functions in @xmath61 are said to be _ integrable _ on @xmath0 . for @xmath62 , @xmath19 , and @xmath67 , the _ lift of @xmath13 at @xmath68 _ , denoted @xmath69 , is the complex valued function of @xmath0 defined by @xmath70 for all @xmath71 . it is easy to see that @xmath61 is the @xmath20-space spanned by @xmath69 for @xmath62 , @xmath19 , @xmath67 . the following is fundamental : [ theorem_main_properties_of_integral ] there is a unique @xmath20-linear functional @xmath21 on @xmath61 which satisfies : 1 . @xmath21 lifts the haar integral on @xmath1 : for @xmath62 , @xmath72 2 . translation invariance : for @xmath64 , @xmath19 , @xmath73 3 . compatibility with multiplicative structure : for @xmath64 , @xmath18 , @xmath74 here the _ absolute value _ of @xmath18 is defined by @xmath75 , and we have adopted the customary integral notation @xmath76 . see @xcite . [ remark_integral_of_lifted_function ] if @xmath69 is as in the previous remark then @xmath77 . [ eg_null_functions ] this example demonstrates some unusual behaviour of the integral . 1 . we will first calculate the measure of @xmath14 . note that @xmath78 , and so @xmath79 2 . for a second example , introduce @xmath80 , and @xmath81 where @xmath82 is a haar measurable subset of @xmath1 of measure @xmath83 ; let @xmath84 . these functions are all complex - valued so it makes sense to consider their absolute values ( here @xmath85 denotes the usual absolute value on @xmath55 , not the absolute value on @xmath0 introduced in the previous theorem ) . the following identities hold : @xmath86 proofs of these identities may be found in @xcite example 1.9 . in this section we briefly explain repeated integration on @xmath22 ; more details are available in @xcite a @xmath20-valued function @xmath8 on @xmath22 is said to be _ fubini _ if and only if both its repeated integrals exist and are equal . that is , we require : 1 . for all @xmath71 , the function @xmath87 is integrable , and then that the function @xmath88 is also integrable ; 2 . for all @xmath89 , the function @xmath90 is integrable , and then that the function @xmath91 is also integrable ; 3 . @xmath92 . similarly , an integrable complex valued function @xmath13 on @xmath93 will be called _ fubini _ if and only if both its repeated integrals exist and are equal . recall that the existence and equality of the repeated integrals of a complex valued function on @xmath93 does not imply its integrability on @xmath93 ( see e.g. ( * ? ? ? * example 8.9c ) ) which is why we have separately imposed that condition . however , in our applications we will restrict to well enough behaved functions for this subtle problem to be irrelevant . fubini s theorem implies that almost all ( in the sense of failing off a set of measure of zero ) the horizontal and vertical sections of any integrable function on @xmath93 are integrable . therefore any integrable function on @xmath93 differs off a null set from some fubini function . however , there is no satisfactory theory of lifting null sets from @xmath59 to @xmath0 , so we restrict attention to fubini functions on @xmath93 . any function in the schwartz - bruhat space of @xmath93 is fubini ; recall that if @xmath59 is archimedean these are the smooth functions of rapid decay at infinity , and if @xmath59 is non - archimedean these are the locally constant functions of compact support . the main properties of the collection of fubini functions on @xmath22 are the following : [ proposition_properties_of_fubini_functions ] the collection of fubini functions on @xmath22 is a @xmath20-space with the following properties : 1 . if @xmath8 is fubini on @xmath22 , then so is @xmath94 for any @xmath95 , @xmath96 , @xmath67 , with repeated integral @xmath97 2 . if @xmath13 is fubini on @xmath93 , then @xmath98 is fubini on @xmath22 , with repeated integral @xmath99 the proof is straightforward ; it may be found in @xcite . [ remark_repeated_integral_of_lifted_function ] the proposition implies that if @xmath13 is fubini on @xmath93 , @xmath100 , @xmath101 , then the function @xmath102 of @xmath22 defined by @xmath103 for all @xmath104 is fubini . the function @xmath8 is said to be the _ lift of @xmath13 at @xmath105_. proposition [ proposition_properties_of_fubini_functions ] implies @xmath106 in this section we investigate the action of polynomials on @xmath0 and the relations with ramification theory ; the results hold for any henselian discrete valuation field @xmath0 with infinite residue field : suppose @xmath107 is a polynomial over @xmath0 , that @xmath32 , @xmath108 are two translated fractional ideals , and that @xmath109 . then there is a unique polynomial @xmath110 $ ] which gives a commutative diagram @xmath111 moreover , @xmath112 . there is certainly at most one function @xmath113 making the diagram commute ; but @xmath114 is an infinite field so if two polynomials are equal as functions then they are equal as polynomials . so there can be at most one polynomial @xmath113 . we may write @xmath115 where @xmath116 $ ] is a polynomial with integer coefficients , no constant term , and with non - zero image in @xmath117 $ ] ( i.e. not all coefficients of @xmath118 are in @xmath119 ) . we clearly have a commutative diagram @xmath120 where @xmath121 denotes the image of @xmath118 in @xmath117 $ ] . if @xmath122 then the inclusion @xmath109 implies @xmath121 is everywhere equal to @xmath123 ; but @xmath114 infinite then implies @xmath121 is a constant polynomial and hence is zero ( since @xmath118 has no constant term ) , which is a contradiction . hence @xmath124 , and we may easily complete the proof : if @xmath125 then the desired commutative diagram is @xmath126 where the lower horizontal map is the function @xmath127 . if @xmath128 then the desired diagram is @xmath129 where the lower horizontal map is the constant function @xmath130 . suppose @xmath107 is a polynomial over @xmath0 , that @xmath32 , @xmath108 are two translated fractional ideals , and that @xmath109 . the unique polynomial @xmath110 $ ] which gives a commutative diagram @xmath111 is said to be the _ residue field approximation _ of @xmath9 with respect to the translated fractional ideals . regarding the previous definition , the translated fractional ideals will usually be clear from the context . the constant term of @xmath113 depends not only the sets @xmath32 and @xmath108 , but on the representatives @xmath131 we choose . when drawing the diagram above , we will henceforth omit the vertical maps , even though they do depend on the choice of @xmath131 . we will also follow the habit used in the previous lemma of denoting a constant function on @xmath114 by the value it assumes . much of this paper is concerned with the problem of explicitly decomposing the preimage of a set under a polynomial and describing the resulting residue field approximations . here is a example to illustrate the ideas : set @xmath132 and assume @xmath133 . the aim of this example is to give explicit descriptions of the sets @xmath134 for @xmath135 , as well as all associated residue field approximations . direct calculations easily show that if @xmath16 , then @xmath136 if and only if @xmath137 or @xmath138 . further , the residue field approximations are @xmath139 similarly , if we suppose @xmath16 then @xmath140 if and only if @xmath141 ; and @xmath142 if and only if @xmath143 . if @xmath114 contains a square root of @xmath144 , let @xmath145 denote a lift of it to @xmath14 ; then @xmath146 with residue field approximations @xmath147 if @xmath114 does not contain a square root of @xmath144 , then @xmath148 , with the residue field approximation given by the third diagram above . we now turn to generalising the example to an arbitrary polynomial ; for later applications to integration the following results will allow us to reduce calculations to the residue field , where we change variable according to the residue field approximation polynomials for example , and then return to @xmath0 . the first decomposition result treats the non - singular part of the polynomial , and is really just a rephrasing of hensel s lemma : [ proposition_non_singular_decomposition ] let @xmath149 be a polynomial with coefficients in @xmath14 , of degree @xmath150 and with non - zero image in @xmath117 $ ] ; let @xmath151 . 1 . suppose that @xmath152 for some @xmath153 and that @xmath154 . then for any @xmath155 , @xmath156 and the residue field approximation is ` multiplication by @xmath157 ' : @xmath158 2 . let @xmath159 be the simple ( i.e. @xmath160 ) solutions in @xmath114 to @xmath161 ; let @xmath162 be a lift by hensel of @xmath163 to @xmath14 ; that is , @xmath164 . then for any @xmath155 , @xmath165 \(i ) is essentially contained in the proof of hensel s lemma and so we omit it . ( ii ) easily follows . we now consider the singular part , which is much more interesting and will be the root of future difficulties : [ proposition_singular_decomposition ] let @xmath149 be a polynomial with coefficients in @xmath14 , of degree @xmath150 and with non - zero image in @xmath117 $ ] ; let @xmath151 . for @xmath155 there is a decomposition @xmath166 ( assuming this set is non - empty i.e. that @xmath167 has a repeated root in @xmath59 ) , where @xmath168 , and @xmath169 are positive integers . first suppose @xmath170 . let @xmath171 be lifts to @xmath14 of the distinct solutions in @xmath59 to @xmath161 and @xmath172 , and set @xmath173 for each @xmath174 . then the required decomposition is @xmath175 we now determine the residue field approximation of @xmath46 on each @xmath176 as it will be used later in corollary [ corollary_r=-1 ] . so , for each @xmath174 , consider the taylor expansion @xmath177 where @xmath178 . but @xmath179 implies @xmath180 for all @xmath181 in @xmath14 , which is to say that @xmath182 commutes , where the lower horizontal map is constant i.e. each residue field approximation associated to the decomposition is constant . we now suppose @xmath183 and proceed by induction on @xmath184 . let @xmath185 be the distinct solutions to @xmath161 and @xmath172 , and write @xmath186 for @xmath187 . since @xmath188 it is enough to decompose each @xmath189 in the required manner , so we now fix a value of @xmath174 , writing @xmath190 and @xmath191 . if @xmath192 is empty then we are done ; else @xmath193 has a lift to @xmath153 such that @xmath194 , and we now fix such an @xmath195 . using the same taylor expansion as above , there exist @xmath196 and @xmath197 $ ] such that @xmath198 and @xmath199 ; in fact , @xmath200 implies @xmath201 , though we will not use this . therefore @xmath202 but also note that @xmath203 by choice of @xmath195 . therefore @xmath204 , and it becomes clear how the induction should proceed . in fact , we must consider three cases , depending on the relative magnitudes of @xmath205 and @xmath184 : 1 . then @xmath207 and @xmath208 ; therefore the residue field approximation is constant , given by the diagram @xmath209 this implies @xmath210 with a constant residue field approximation : @xmath211 2 . again , @xmath207 ; the residue field approximation is clearly @xmath213 therefore @xmath210 , with residue field approximation @xmath214 3 . @xmath215 . here we may use the inductive hypothesis and proposition [ proposition_non_singular_decomposition ] to write @xmath216 with residue field approximations @xmath217 , say : @xmath218 therefore @xmath219 , with residue field approximations @xmath220 for @xmath149 as in the previous two propositions , these two decomposition results completely describe @xmath221 in terms of @xmath222 translated fractional ideals equipped with polynomial residue field approximations . moreover , the proof of the second result gives some insight into how structure of the polynomial @xmath46 effects the resulting residue field approximations . for applications beyond those described in this paper , it will be necessary to better understand how the decomposition varies with @xmath223 and @xmath184 . for small @xmath184 we have the following result : [ lemma_depth_1_and_2 ] let @xmath149 be a polynomial with coefficients in @xmath14 , of degree @xmath150 and such that @xmath224 has non - zero image in @xmath117 $ ] ; let @xmath170 or @xmath225 . there are finitely many @xmath226 such that if @xmath227 and @xmath228 is non - empty , then @xmath229 for some @xmath230 . first suppose @xmath170 . then @xmath231 being non - empty implies that @xmath232 is the image under @xmath45 of one of the finitely many roots of @xmath233 . now suppose that @xmath234 . then the argument is just the same as for @xmath170 , except it is important to observe the following : if @xmath235 are equal modulo @xmath119 , and @xmath236 for @xmath237 , then @xmath238 . this follows from the taylor expansion and the fact that @xmath239 . decomposition results similar to the previous ones are common in model theory ; for example , in the theory of algebraically closed valued fields @xcite , every definable subset of the field is a finite disjoint union of points and ` swiss cheeses ' . further , these decompositions are related to ramification theory and rigid geometry through a. abbes and t. saito s @xcite @xcite ramification theory for complete discrete valuation fields with imperfect residue field , about which we will now say some more . until abbes and saito work it was a significant open problem to systematically generalise the classical ramification theory for complete discrete valuation fields with perfect residue field to the imperfect residue field situation ; some alternative approaches are due to j. borger @xcite @xcite , k. kato @xcite @xcite , and i. zhukov @xcite @xcite . geometrically , the importance of this lies in the following situation . if @xmath240 is a finite morphism between smooth , projective surfaces , over a field @xmath241 which is allowed to be perfect , then according to zariski s ` purity of the branch locus ' , the ramification of @xmath242 occurs along curves . let @xmath243 be an irreducible curve with generic point @xmath244 , and set @xmath245 this is a complete discrete valuation field whose residue field is @xmath246 . moreover , we have a finite extension @xmath247 whose ramification properties reflect the local ramification of @xmath242 along @xmath248 . but @xmath246 will be imperfect and @xmath249 may have an inseparable residue field extension . we now give a summary of the basics of abbes and saito s theory . there is a more extensive overview by l. xiao @xcite . let @xmath250 be a finite , galois extension of complete discrete valuation fields with arbitrary residue fields . then @xmath251 is a complete intersection algebra over @xmath252 ( since they are both regular local rings ) and we may therefore write @xmath253/\langle f_1,\dots , f_n\rangle\ ] ] for a regular sequence @xmath254 . now , for any real @xmath255 , one introduces the rigid space @xmath256 where @xmath257 is the extension of the discrete valuation on @xmath258 . let @xmath259 denote the set of connected components ( in the sense of rigid geometry ) of this space ; if @xmath260 , which one may assume if @xmath261 is separable , then it follows from our decomposition results [ proposition_non_singular_decomposition ] and [ proposition_singular_decomposition ] that these components are balls . as @xmath262 , @xmath263 will consist of @xmath264 small balls ; conversely , @xmath265 is a single large ball . a central idea of abbes and saito s theory is to analyse the behaviour of @xmath263 as @xmath195 varies ; in particular , when it breaks into @xmath266 balls . this will soon be made precise . the natural action of the absolute galois group @xmath267 on @xmath263 induces an action on @xmath259 , which then factors transitively through @xmath268 . [ remark_ramification_theory_in_perfect_case ] to motivate what follows , let us briefly suppose that @xmath258 has perfect residue field . let @xmath269\to[-1,\infty]$ ] be the inverse of the hasse - herbrand function @xmath270 , relating the upper and lower ramification filtrations by @xmath271 ( see @xcite or @xcite ) . then it is not hard to prove : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for @xmath272 , @xmath273 acts trivially on @xmath274 if and only if @xmath275 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ( a nice sketch is given in @xcite ) . so , for any @xmath272 , the kernel of the action of @xmath276 on @xmath277 is @xmath278 . abbes and saito take this final observation in this remark as the definition of the upper filtration in their theory : let @xmath250 be a finite , galois extension of complete discrete valuation fields . the _ upper ramification filtration _ on @xmath268 , is defined , for @xmath272 , by @xmath279 starting from this definition of the upper ramification filtration , abbes and saito develop fully a ramification theory for @xmath258 . xiao has augmented their work by establishing the hasse - arf integrality theorem for certain conductors @xcite @xcite . now suppose that @xmath280 is our two - dimensional local field . as discussed above , a problem which will appear later is the variation of the sets @xmath221 as @xmath223 varies across @xmath281 . this is related , albeit mysteriously , with understanding the ramification of extensions of the form @xmath282 , where @xmath283 is a root of @xmath284 . in fact , we will soon assume that @xmath285 , and an alternative viewpoint is that both the integration and ramification theories study the monodromy action of @xmath286 around zero . unfortunately , there are currently no precise comparison results between the integration and ramification theories . in this section we investigate the behaviour of fubini functions on @xmath22 under certain non - linear changes of variables . more precisely , we consider the following : [ conjecture ] let @xmath100 , @xmath101 , and let @xmath107 be a polynomial over @xmath0 . then for any schwartz - bruhat function @xmath13 on @xmath93 , letting @xmath102 be the lift of @xmath13 at @xmath105 , the function @xmath287 is fubini on @xmath22 , with repeated integral equal to that of @xmath13 . the conjecture is false in the generality in which we have stated it , though an important special case has already been done treated in a previous paper of the author : [ theorem_linear_case ] with notation as in the conjecture , if @xmath288 then the conjecture is true . the main result of @xcite , for a two - dimensional space , is that for any @xmath289 , the function @xmath290 is fubini on @xmath22 . the conjecture is a special case of that result when @xmath291 ; in fact , it essentially follows from lemma 3.7 of @xcite if @xmath292 then the conjecture follows from translation invariance of the integral ; see proposition [ proposition_properties_of_fubini_functions ] and remark [ remark_repeated_integral_of_lifted_function ] . because of the previous theorem , we will have in mind polynomials @xmath107 of degree at least @xmath225 , though our results are equally valid for lower degree . we will be interested in conditions on the data @xmath293 such that the conjecture is true for all schwartz - bruhat functions @xmath13 . we assign to the data two invariants as follows : let @xmath293 be data for the conjecture , and write @xmath294 , where @xmath295 , @xmath40 $ ] , and the image of @xmath46 in @xmath117 $ ] is non - zero . note that @xmath285 . the _ depth _ and _ normalised polynomial _ associated to the data are defined to be @xmath296 and @xmath149 respectively . a summary of what we know about the validity of the conjecture , classified by the depth and normalised polynomial , may be found in section [ section_summary ] . the sense in which the depth and normalised polynomial are invariants , and why they are useful , is given by the following lemma in which we reduce the conjecture to a special case : [ lemma_reduction ] fix a polynomial @xmath40 $ ] with nonzero image in @xmath117 $ ] and no constant term , and an integer @xmath295 . then the following are equivalent : 1 . the conjecture is true for all data @xmath293 with depth @xmath297 and normalised polynomial @xmath46 ; 2 . the conjecture is true for all data of the form @xmath298 with depth @xmath297 , normalised polynomial @xmath46 , and such that @xmath299 ; 3 . for all schwartz - bruhat functions @xmath13 on @xmath93 , the function @xmath300 is fubini ; 4 . for all schwartz - bruhat functions @xmath13 on @xmath93 , the following hold : for each @xmath89 , the function @xmath301 is integrable , then that @xmath302 is integrable , and finally that @xmath303 clearly ( i)@xmath304(ii ) . the only data satisfying the conditions of ( ii ) are @xmath305 , and so ( ii)@xmath306(iii ) . ( iii)@xmath304(i ) : so assume ( iii ) , letting @xmath293 be data for the conjecture with depth @xmath297 and normalised polynomial @xmath46 . let @xmath13 be schwartz - bruhat on @xmath93 and write @xmath102 . note that @xmath307 , and that therefore for all @xmath104 , @xmath308 by ( iii ) , this final function of @xmath309 differs from a fubini function by translation . so @xmath310 differs from a fubini function only by translation and scaling , and hence is itself fubini , by proposition [ proposition_properties_of_fubini_functions ] . therefore we have proved ( i ) . ( iii)@xmath306(iv ) : first note that for any @xmath71 , the function @xmath311 is just the translation of @xmath312 by @xmath313 ; since @xmath12 is fubini this is integrable , and translation invariance of the integral implies @xmath314 but as a function of @xmath181 this is integrable , again since @xmath12 is fubini , and @xmath315 now by remark [ remark_repeated_integral_of_lifted_function ] and fubini s theorem for @xmath93 , @xmath316 by the definition of a fubini function , it now follows that @xmath317 is fubini if and only if the @xmath318 repeated integral is well - defined and equals @xmath319 , which is precisely what is stated in ( iv ) . with these reductions at hand it is straightforward to establish the conjecture in the case of non - negative depth : [ theorem_r_positive ] let @xmath293 be data for the conjecture , and suppose that the associated depth is non - negative . then the conjecture is true . by the reductions , we suppose that @xmath40 $ ] is a polynomial with no constant term and non - zero image in @xmath117 $ ] , that @xmath320 is an integer , and we will prove condition ( iv ) of the lemma above . write @xmath321 , and let @xmath13 be schwartz - bruhat on @xmath93 . the assumption on @xmath297 implies that all coefficients of @xmath9 are integral , and for @xmath89 we have @xmath322 hence if @xmath323 , we see that @xmath324 is the lift of @xmath325 at @xmath326 , where @xmath327 is the image of @xmath9 in @xmath117 $ ] . if @xmath328 , then @xmath329 for all @xmath181 in @xmath0 . integrating with respect to @xmath181 therefore obtains @xmath330 which simply says that @xmath331 is the lift of @xmath332 at @xmath326 . hence we may integrate with respect to @xmath244 to get @xmath333 where the second line follows from the first by fubini s theorem on @xmath93 . the result now follows by translation invariance of the measure on @xmath59 and lemma [ lemma_reduction ] . this is convenient opportunity for an additional remark on ramification theory . according to our discussion at the end of section [ section_decompositions_and_ramificiation ] , the conjecture relates to the monodromy or ramification of @xmath334 around the origin . non - negative depth corresponds to the unramified situation , so the previous theorem could be stated as `` the conjecture is true in the unramified case '' . having reduced the problem as far as possible and treated the relatively easy case , we discuss the case of negative depth in this section and the following section [ section_purely_insep ] . for this section and the next we fix the following notation : @xmath38 a negative integer as the depth ; a polynomial @xmath40 $ ] without constant term and with non - zero image in @xmath117 $ ] as the normalised polynomial ; and a schwartz - bruhat function @xmath13 on @xmath93 . write @xmath335 for the function of @xmath22 given by @xmath336 , and @xmath45 for the image of @xmath46 in @xmath117 $ ] . in this section , we also assume that @xmath45 does not have everywhere vanishing derivative ; since @xmath45 is non - zero and without constant term , this condition can only fail to be satisfied if @xmath114 has positive characteristic @xmath337 and @xmath149 is a purely inseparable polynomial i.e. a polynomial in @xmath338 . we shall drop this assumption in section [ section_purely_insep ] and see that conjecture [ conjecture ] fails for such highly singular @xmath46 . we will study the conjecture for data of depth @xmath297 and normalised polynomial @xmath46 through condition ( iv ) of lemma [ lemma_reduction ] . we will establish various conditions under which the conjecture holds . [ remark_unferocious_case ] from the point of view of monodromy , we are now in the unferocious case ; recall that an extension of complete discrete valuation fields is said to be _ unferocious _ so long as the residue field extension is separable . the author believes that this terminology is due to i. zhukov @xcite . introduce two sets : the non - singular part of @xmath46 @xmath339 and the singular part @xmath340 by our assumption on @xmath46 , the non - singular part @xmath341 is non - empty . the corresponding singular and non - singular parts of @xmath335 are the restriction of @xmath335 to these sets extended by zero elsewhere : @xmath342 note that @xmath343 . the singular and non - singular parts are treated separately . using the decomposition result [ proposition_non_singular_decomposition ] , we will now explicitly evaluate @xmath344 for any @xmath89 : for all @xmath89 , the function @xmath344 is integrable , and @xmath345 is the lift of @xmath346 at @xmath347 ; the sum is taken over all simple solutions @xmath348 to @xmath349 . moreover , this function @xmath345 is integrable on @xmath0 , with @xmath350 firstly , for @xmath351 , we have @xmath352 for all @xmath71 . now fix @xmath353 . then for @xmath354 to be non - zero , @xmath181 must belong to @xmath355 where @xmath162 are lifts by hensel of the simple solutions @xmath163 in @xmath114 to @xmath356 and the decomposition is provided by the decomposition result [ proposition_non_singular_decomposition ] ; that proposition also implies that there are commutative diagrams @xmath357 so we write @xmath358 , where @xmath359 is the restriction of @xmath344 to @xmath360 , extended by zero elsewhere ; if @xmath361 belongs to @xmath360 then the commutative diagram implies @xmath362 therefore @xmath359 is the lift of the haar integrable function @xmath363 at @xmath364 , the integral of which is @xmath365 by remark [ remark_integral_of_lifted_function ] . by linearity , @xmath344 is integrable , with @xmath366 the previous paragraph considered a fixed value of @xmath367 in @xmath368 . we now consider the integral ( @xmath369 ) as a function of @xmath244 ; that is , @xmath370 recall that @xmath159 are the simple solutions in @xmath59 to @xmath356 . so we may rewrite the integral as @xmath371 where the sum is over the finitely many @xmath348 in @xmath59 which satisfy @xmath372 and @xmath373 . finally , by the appendix , the function @xmath374 is in fact haar integrable on @xmath59 with integral @xmath375 therefore @xmath345 is integrable on @xmath0 , with @xmath350 the proposition has an immediate corollary : [ corollary_nowhere_singular ] if @xmath376 is no - where vanishing on @xmath114 , then @xmath335 is fubini . if @xmath376 has no roots in @xmath114 , then @xmath377 , so the previous proposition and lemma [ lemma_reduction ] imply @xmath335 is fubini . more generally , the proposition reduces the problem to showing that the singularities of @xmath45 give no contribution to the integrals : [ corollary_reduction_to_singular_part ] the function @xmath335 is fubini if and only if the following hold : for each @xmath89 , the function @xmath378 is integrable , then that @xmath379 is integrable , and finally that @xmath380 this follows immediately from the identity @xmath343 , the previous proposition , lemma [ lemma_reduction ] , and linearity . we may verify the first requirement of corollary [ corollary_reduction_to_singular_part ] using the decomposition result [ proposition_singular_decomposition ] : [ proposition_integral_of_singular_sections ] for each @xmath89 , the function @xmath378 is integrable , and we have the following explicit descriptions of its integral : if @xmath351 , or if @xmath381 but @xmath382 is empty , then @xmath383 . otherwise we have @xmath381 and write @xmath384 where the decomposition ( which depends on @xmath244 ) is provided by the decomposition result [ proposition_singular_decomposition ] ; let @xmath385 $ ] for @xmath187 denote the corresponding residue field actions i.e. @xmath386 commutes . then @xmath387 where the summation @xmath388 is over those @xmath389 for which @xmath390 is not a constant polynomial . by the definition of a lifted function , @xmath12 vanishes off @xmath24 . so if @xmath382 is empty for some @xmath244 then @xmath378 is everywhere zero and hence integrable ; note that this set is certainly empty if @xmath351 . now fix @xmath353 for the remainder of the proof . then for @xmath71 , @xmath391 vanishes unless @xmath181 belongs to @xmath392 where the decomposition is as in the statement of the proposition ; let @xmath390 be the corresponding residue field approximations . denote by @xmath393 the restriction of @xmath378 to @xmath176 , extended by zero elsewhere . we shall now prove that each @xmath393 is an integrable function . indeed , @xmath393 vanishes off @xmath176 , and if @xmath394 , then @xmath395 by definition of the residue field approximation @xmath390 . therefore @xmath393 is a lifted function : it is the lift of @xmath396 at @xmath397 . further , since we assumed @xmath13 is schwartz - bruhat , this function of @xmath193 is haar integrable on @xmath59 _ so long as @xmath390 is not constant _ , and therefore @xmath393 is integrable on @xmath0 , with @xmath398 however , if @xmath390 is a constant polynomial , then @xmath399 , which is integrable with zero integral by example [ eg_null_functions ] . by linearity , @xmath378 is integrable , with @xmath387 as required . we emphasise again that the decomposition @xmath400 which we have used to express the integral _ depends on @xmath244_. [ corollary_r=-1 ] if @xmath44 then @xmath335 is fubini . looking at the proof of decomposition result [ proposition_singular_decomposition ] , we see that if @xmath44 ( i.e. @xmath170 in the notation of that result ) , then all the residue field approximations are constant . so by the previous proposition , @xmath383 for all @xmath89 . corollary [ corollary_reduction_to_singular_part ] implies @xmath335 is fubini . by proposition [ proposition_integral_of_singular_sections ] we now have a well defined function @xmath379 ; to establish the validity of the conditions of corollary [ corollary_reduction_to_singular_part ] the next step is to prove that this function of @xmath244 is integrable . the complication in establishing its integrability is that we lack explicit information on the variation of the sets @xmath401 as @xmath402 runs though @xmath14 . we now present some results and calculations which reveal considerable insight into why @xmath379 can in fact fail to be integrable . we shall also give evidence that this phenomenon is merely a result of the integration theory not yet being sufficiently developed . [ proposition_finite_image ] assume that there exist @xmath226 such that if @xmath227 and @xmath403 is non - empty , then @xmath404 for some @xmath230 . note that this is satisfied if @xmath44 or @xmath405 , by corollary [ lemma_depth_1_and_2 ] . then @xmath379 is a finite sum of lifts of functions of the form @xmath406 for @xmath407 $ ] non - constant , @xmath408 , and @xmath409 . let @xmath410 be as in the statement of the proposition ; we also assume that @xmath410 are distinct modulo @xmath411 . by proposition [ proposition_integral_of_singular_sections ] , if @xmath89 is not in @xmath412 for some @xmath145 , then @xmath383 . so letting @xmath413 be the restriction of @xmath379 to @xmath412 , extended by zero elsewhere , we have an equality of functions of @xmath244 : @xmath414 for convenience of notation , we now fix some @xmath145 and write @xmath415 , @xmath416 . write @xmath417 with residue field approximations @xmath390 . we claim that @xmath276 is the lift of @xmath418 at @xmath419 ( the sum @xmath420 is restricted to those @xmath174 such that @xmath390 is not constant ) . so suppose @xmath421 . then of course @xmath422 , and so @xmath423 with the residue field approximations of this decomposition given by @xmath424 proposition [ proposition_integral_of_singular_sections ] implies @xmath425 proving the claim , and completing the proof . [ remark_nonsense ] suppose that the assumption of the previous proposition is satisfied . then to establish integrability of @xmath379 and prove it has zero integral , it is enough to prove that for any @xmath408 , non - constant @xmath407 $ ] , and @xmath8 schwartz - bruhat on @xmath59 , the lift of @xmath426 at @xmath326 is integrable and has zero integral ; let @xmath276 denote this function of @xmath0 , that is , @xmath427 _ then @xmath276 may not be integrable on @xmath0_. indeed , it is not hard to show that if @xmath276 were to belong to @xmath61 , the space of integrable functions , then @xmath276 would be the lift at @xmath326 of a haar integrable function on @xmath59 ; this haar integrable function would then have to be @xmath426 , but the arguments to follow reveal that this function is haar integrable if and only if @xmath428 . we now offer the following nonsense argument for why @xmath276 should be integrable , and why @xmath429 should be zero . as a lifted function , we evaluate the integral of @xmath276 by theorem [ theorem_main_properties_of_integral ] to give @xmath430 and then apply fubini s theorem for @xmath59 and translation invariance of the integral to deduce @xmath431 at this point it is clear why our arguments are not valid : the function @xmath426 is not integrable on @xmath59 . however , we may apply similar nonsense to the function @xmath432 , which is the lift of @xmath433 , to deduce @xmath434 finally , example [ eg_null_functions](i ) implies @xmath435 and so @xmath436 it should be possible to extend the measure theory on @xmath0 so that these manipulations become rigorous . the key idea is that from the vantage point of @xmath0 , the residue field @xmath59 truly has zero measure , as used above ; so one expects fubini s theorem on @xmath59 to hold for certain functions which , though not haar integrable , are integrable in some sense after imposing the condition @xmath437 . once this is properly incorporated into the measure , the theory should become considerably richer . it should also yield new methods to treat divergent integrals on @xmath59 by lifting them to @xmath0 , applying fubini theorem there , and then pulling the results back down to @xmath59 ; this would be a refreshing contrast to the main techniques so far , which have centred around reducing integrals on @xmath0 down to @xmath59 . [ example_nonsense ] now we treat an example of depth @xmath438 in which the assumption of proposition [ proposition_finite_image ] is not satisfied . we assume @xmath439 , @xmath440 , and @xmath133 . the decompositions required for the proposition are given by @xmath441 where we use hensel s lemma to take a square root in the third case . the associated residue field approximations in the final two cases are given by @xmath442 proposition [ proposition_integral_of_singular_sections ] therefore implies that for @xmath89 , @xmath443 therefore @xmath444 is the lift of @xmath445 at @xmath446 . this function of @xmath0 need not be integrable , but as in the previous remark , there is a good argument to suggest that it should be , and why its integral should be zero : indeed , the function on the residue field has the form @xmath447 where @xmath8 is a schwartz - bruhat function on @xmath59 . now replace the integrand by @xmath448 and appeal to the appendix to deduce @xmath449 but arguing as in the proceeding remark , @xmath450 , and so @xmath451 . of course , the argument is nonsense because @xmath452 is not integrable , but it should be after a suitable extension of the measure . we maintain all notation introduced at the beginning of the previous section but drop the additional hypothesis that @xmath233 is not everywhere zero . instead , we now assume @xmath59 has positive characteristic @xmath337 and that @xmath453 is purely inseparable . whereas in the previous section conjecture [ conjecture ] could fail to hold because the integration theory is not yet sufficiently developed , causing functions not to be integrable , we will present a result now to show that if @xmath45 is purely inseparable then all required functions are integrable , but the conjecture is simply false ! first note that , in the notation of the previous section , @xmath233 being everywhere zero implies @xmath454 . secondly , proposition [ proposition_integral_of_singular_sections ] remains valid , so that @xmath455 is integrable for any @xmath89 and we have an explicit description of its integral . [ proposition_r=-1_purely_insep ] suppose @xmath44 . then both repeated integrals of @xmath335 are well - defined , but @xmath13 may be chosen so that @xmath456 arguing exactly as in corollary [ corollary_r=-1 ] it follows that @xmath383 for all @xmath89 , and therefore @xmath457 is certainly integrable , with integral @xmath458 . that is , @xmath459 the @xmath460 integral of @xmath335 was showed to make sense in lemma [ lemma_reduction ] and have value @xmath461 to complete the proof simply choose @xmath13 to be any schwartz - bruhat function on @xmath93 such that @xmath319 is non - zero . of course , using the terminology of [ remark_unferocious_case ] , we are essentially now studying the ramification of ferocious extensions . further , the integration theory of @xcite is easily modified to allow integration on a complete discrete valuation field whose residue field is any infinite field equipped with discrete counting measure ; this is an elementary form of motivic integration . in that situation one may ask similar question about changes of variables and fubini s theorem ; results are generally easier to prove and closer to the analogous results for a usual local field . if the residue field is perfect , then the pathologies exhibited in this section no longer exist . perhaps the most immediate relation between measure theory and imperfectness is that the set of @xmath462 powers of @xmath59 have zero measure , in stark contrast with in a perfect field . let us summarise our main results on conjecture [ conjecture ] . given data @xmath293 for the conjecture , let @xmath46 be the associated normalised polynomial and @xmath297 the depth ; then : 1 . if @xmath463 then the conjecture is true ( theorem [ theorem_linear_case ] ) . 2 . if @xmath464 then the conjecture is true ( theorem [ theorem_r_positive ] ) 3 . if @xmath233 is no - where vanishing on @xmath114 then the conjecture is true ( corollary [ corollary_nowhere_singular ] ) . 4 . if @xmath44 and @xmath45 is not purely inseparable , then the conjecture is true ( lemma [ lemma_reduction ] @xmath465 corollary [ corollary_r=-1 ] ) . if @xmath48 and @xmath45 is not purely inseparable , then @xmath466 may fail to be integrable and so the conjecture may fail ; it appears that it is possible to increase the space of integrable functions so that the conjecture becomes true ( remark [ remark_nonsense ] @xmath465 example [ example_nonsense ] ) . 6 . if @xmath44 but @xmath45 is purely inseparable , then the conjecture fails and would continue to fail even if we increased the scope of the integral ( section [ section_purely_insep ] ) . 7 . if @xmath48 but @xmath45 is purely inseparable , then similarly to case ( v ) calculations become difficult . we have not included examples , but in all cases which the author can explicitly evaluate , @xmath467 , thus the conjecture seems to fail as in ( vi ) . there are some immediate directions which demand study : 1 . the integral must be extended to a wider class of functions so that the nonsense manipulations in remark [ remark_nonsense ] and example [ example_nonsense ] become valid . we should consider more general changes of coordinates on @xmath22 than @xmath10 . similar methods to those in this paper will be required : firstly one needs to approximate the transformation at the level of @xmath93 and find a suitable decomposition . this will lead to integrals over @xmath59 which can be explicitly evaluated as well as some functions on @xmath0 ; these functions on @xmath0 will perhaps be within the scope of the integral , or instead will provide further impetus for extending the integral . 2 . the second issue concerns ramification . the imprecise relations between integration and ramification ( it seems to be more useful to take the point of view of monodromy ) which we have had occasion to mention should be made concrete . perhaps this will be possible through a suitable theory of two - dimensional local zeta functions , just as the one - dimensional theories of local zeta functions and igusa zeta functions capture ramification data of abelian extensions and monodromy of singularities . an enjoyable introduction to some of these topics , in dimension one , may be found in @xcite also , the following seems worth consideration : let us return to the situation of remark [ remark_ramification_theory_in_perfect_case ] , supposing that @xmath250 is a finite , galois extension of complete , discrete valuation fields with perfect residue fields . the definition of the hasse - herbrand function implies that @xmath468 at least away from the ramification breaks , and therefore that @xmath469 since both sides are @xmath470 at @xmath471 . but @xmath472 , and so @xmath473 for all @xmath272 . if we think of ` the number of connected components ' as a measure , then ( @xmath369 ) is a repeated integral taken over certain fibres , and it is exactly the variation of the fibres which contributes to the interesting structure of the hasse - herbrand function . whether this repeated integral interpretation of ramification can be more systemically exploited in the local setting is an interesting question . let @xmath59 be a local field , @xmath13 a fubini function of @xmath93 , and @xmath407 $ ] a polynomial with @xmath474 not everywhere zero . we discuss the function of @xmath59 given by @xmath475 note that the assumption that @xmath13 is fubini implies that @xmath452 is defined ( i.e. not infinite ) for all @xmath476 . we will prove the following : let @xmath479 and assume that there is a non - singular solution to @xmath480 . the inverse function theorem for complete fields ( see e.g. @xcite ) implies that there exists an open disc @xmath481 , open discs @xmath482 , and @xmath59-analytic maps @xmath483 , @xmath484 ( that is , representable in @xmath485 by a convergent power series ) with the following properties : 1 . @xmath482 are pairwise disjoint ; 2 . @xmath486 for each @xmath145 ; moreover , @xmath487 and @xmath488 are inverse diffeomorphisms between @xmath489 and @xmath485 ; 3 . the non - singular solutions in @xmath59 to @xmath480 are @xmath490 . for if not , then there would exist a sequence @xmath494 in @xmath59 such that @xmath495 for all @xmath496 and @xmath497 ; the relative compactness of @xmath498 now allows us to pass to a convergent subsequence of @xmath499 , giving an element @xmath500 which satisfies @xmath501 . but this contradicts ( iii ) and so proves our claim . informally , the @xmath488 parametrise the non - singular solutions of @xmath492 , for @xmath491 . for @xmath491 , we deduce that @xmath502 and so @xmath503 by fubini s theorem and an analytic change of variables @xmath504 . an elementary introduction to change of variables in integrals over non - archimedean fields may be found in @xcite . it is now easy to see that @xmath508 admits a partition into countably many borel sets @xmath509 where ( @xmath369 ) holds with @xmath510 in place of @xmath184 for each @xmath184 . therefore @xmath511 where @xmath512 differs from @xmath59 only by a finite set . so we have reached the desired result : @xmath513 , ` local fields and their extensions ' , volume 121 of _ translations of mathematical monographs_. american mathematical society , providence , ri , second edition , 2002 . with a foreword by i. r. shafarevich . , ` measure , integration and elements of harmonic analysis on generalized loop spaces ' , in proceedings of the st . petersburg mathematical society . xii , vol . 219 of amer 2 , providence , ri , 2006 , amer . math . soc . , pp . 149165 . , ` swan conductors for characters of degree one in the imperfect residue field case ' . in _ algebraic @xmath59-theory and algebraic number theory ( honolulu , hi , 1987 ) _ , volume 83 of _ contemp . math . _ , pages 101131 . soc . , providence , ri , 1989 . , ` algebraic number theory ' , volume 322 of _ grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences]_. springer - verlag , berlin , 1999 . translated from the 1992 german original and with a note by norbert schappacher , with a foreword by g. harder .

we consider non - linear changes of variables and fubini s theorem for certain integrals over a two - dimensional local field . an interesting example is presented in which imperfectness of a positive characteristic local field causes fubini s theorem to unexpectedly fail . the relationship to ramification theory is discussed .

one of the oldest problem in galactic dynamics is the determination of the mass distribution in the stellar and gaseous disk from observed velocities . observationally , noisy measurements along with a certain velocity dispersion makes this inverse problem difficult to solve ( beauvais & bothum , 2001 ) . from a theoretical point of view , various levels of approximation are possible . for instance , the inversion is straightforward if the disk is considered as a flattened spheroid , since the surface density is simply obtained by an inverse abel transform over the velocity field ( brandt , 1960 ) . flattened spheroids are however a poor representation of real disks ( e.g. kochanek 2001 ) . another , more intuitive way to proceed is to generate a collection of surface density profiles , then to determine the corresponding potential or gravitational field , and find only those which model the velocity data at best . for a razor - thin axi - symmetrical disk , only one numerical quadrature over its radial extent must be performed to obtain the potential or the field , which is a priori very convenient because time in - expansive . meanwhile , this operation is commonly believed to be numerically uncomfortable when using elliptic integrals ( binney & tremaine , 1987 , cuddeford , 1993 , kochanek , 2001 ) . the reason is that these functions contain a singularity everywhere in the source . as suggested by binney & tremaine ( 1987 ) , this drawback can be simply avoided by considering field points located just above / below the equatorial plane . even if the potential and its radial derivative do generally not exhibit strong variations slightly off the mid - plane , such a vertical shift introduces an inconsistency or a bias . as a matter of fact , instead of elliptic integrals , most authors work with bessel functions which have finite amplitudes ( toomre , 1963 ) . but their oscillatory behavior and specially their spatial extension ( the infinite range ) are a severe disadvantage from a numerical point of view ( cuddeford , 1993 ; cohl & tohline 1999 ) . + in this paper , we show that point mass singularities can be properly and easily handled numerically from elliptic integrals by density splitting . for the case of a disk with zero thickness as considered here , the gravitational potential and field can be determined exactly in the equatorial plane without any shift , and whatever the surface density profile ( provided it vanishes at the inner and outer edges of the disk ) . the extension to tri - dimensional disks is possible . for a disk with zero thickness , the expression for the radial gravitational field in the equatorial plane is known in a closed form . using cylindrical coordinates , this reads ( e.g durand , 1964 ) : @xmath0da , \label{eq : potential}\ ] ] where @xmath1 is the total surface density , @xmath2 is the field point , @xmath3 refers to the material distribution , @xmath4 and @xmath5 are respectively the inner and the outer radius of the disk , and @xmath6 and @xmath7 are respectively defined by @xmath8 and @xmath9 finally , @xmath10 ( respectively @xmath11 ) is the complete elliptic integral of the first ( resp . second ) kind . note that the integrand of equation ( [ eq : potential ] ) diverges as the modulus @xmath12 , namely when @xmath13 . the nature of the singularity is twofold : a logarithmic singularity due to the @xmath10-function , and a hyperbolic one due to the term @xmath14 . however , this singularity is integrable . there are different ways to estimate improper integrals . here is a simple and efficient recipe . let us write the surface density as @xmath15 where @xmath16 is the surface density at the field point and @xmath17 is the `` residual '' surface density which , by construction , equals zero when @xmath18 . thus , equation ( [ eq : potential ] ) can be written as the sum of two contributions , namely : @xmath19 where @xmath20da \label{eq : hom}\ ] ] is the radial gravitational field due to a radially homogeneous disk and @xmath21da \label{eq:1}\ ] ] corresponds to the residual profile . the point is that the expression for @xmath22 also exists in a closed analytical form ( see appendix [ sec : appena ] ) . this quantity is finite everywhere inside the disk , even at the disc edges provided that @xmath16 vanishes continuously there . it is important to mention that this restriction is no more necessary if we consider the disk as a tri - dimensional system . besides , @xmath23 can be easily computed numerically because both @xmath24 and @xmath25 are fully regular whatever @xmath3 and @xmath2 , and especially for @xmath18 ( see appendix [ sec : appenb ] for a proof ) . as a consequence , the accuracy on @xmath23 ( and subsequently on the total field @xmath26 ) depends only on the performance of the quadrature scheme used to performed the radial integral in equation ( [ eq:1 ] ) . in figure [ figure ] , we illustrate this simple and efficient technique by considering a disk with inner edge @xmath27 , outer edge @xmath28 and surface density @xmath29 as commonly used to model spiral galaxies ( freeman , 1970 ) . the graph shows the residual profile @xmath30 , as well as the two regular functions @xmath24 and @xmath25 . here , the field point where stands the singularity is arbitrarily set to @xmath31 ( about the middle of the disk ) . we can now apply this method to compute the rotation law @xmath32 of any disk , whatever its size and surface density profile . here , we simply assume that the contribution to the rotation is only gravitation ( and neglect pressure and viscosity effects ) . thus , @xmath32 is defined by the relation @xmath33 where : @xmath34 and @xmath35 using the previous expressions for the radial field , these two terms respectively read : @xmath36-{\bf e}\left(\frac{r_{\rm in}}{r}\right)+{\bf k}\left(\frac{r_{\rm in}}{r}\right)\right\ } \label{eq : omegahomo}\ ] ] and @xmath37da \label{eq:2}\ ] ] these two expressions are easy to compute numerically , for reasons mentioned above . let us remind that at the edges , @xmath38 vanishes because @xmath39 as assumed ( see above ) . we have reported a simple and efficient method to compute the radial field and the rotation curve of any razor - thin axi - symmetric disk in the mid - plane , using complete elliptic integrals . as a matter of fact , a similar treatment can be considered for the gravitational potential . even , this splitting technique can be employed to disks with finite ( non - zero ) thickness as suggested in hur ( 2003 ) , and to the fully tri - dimensional systems as well ( pierens & hur 2004 ) . abramowitz , m. & stegun , i. a. 1970 , handbook of mathematical functions , new york : dover , 1970 beauvais , c. & bothun , g. 2001 , , 136 , 41 binney , j. & tremaine , s. 1987 , princeton , nj , princeton university press , 1987 , 747 p. brandt , j. c. 1960 , , 131 , 293 cohl , h. s. & tohline , j. e. 1999 , , 527 , 86 durand e. , 1964 , in _ `` electrostatique i. les distributions '' _ , masson ed . freeman , k. c. 1970 , , 160 , 811 gradshteyn , i. s. & ryzhik , i. m. 1980 , new york : academic press , 1980 . hur j - m . , 2003 , submitted to pierens a. , & hur j.m . , 2004 , in preparation kochanek c.s . , 2001 , submitted ( astro - ph/0108162 ) toomre , a. 1963 , , 138 , 385 sofue , y. & rubin , v. 2001 , , 39 , 137 let us begin with the expression for the radial field due to a disk with constant surface density , i.e. equation ( [ eq : hom ] ) . this expression can be written as follows : @xmath40da+\int_{r}^{r_{\rm out}}\sqrt\frac{a}{r}k\left[\frac{{\bf e}(k)}{\varpi}-{\bf k}(k)\right]da\right\ } \label{eq : grhom}\ ] ] we can now set @xmath41 in the first integral and @xmath42 in the second one . changing the modulus in the elliptic integrals according to ( gradshteyn & ryzhik , 1980 ) : @xmath43 and @xmath44\ ] ] with @xmath45 and @xmath46 , then equation ( [ eq : grhom ] ) becomes : @xmath47 \label{eq : grhom2}\ ] ] each integral in this expression is known . actually , we have ( gradshteyn & ryzhik , 1980 ) : @xmath48 and @xmath49 inserting these relations into equation ( [ eq : grhom2 ] ) and re - arranging terms yields the formula @xmath50-{\bf e}\left(\frac{r_{\rm in}}{r}\right)+{\bf k}\left(\frac{r_{\rm in}}{r}\right)\right\}\ ] ] in order to compute numerically equations ( [ eq:1 ] ) and ( [ eq:2 ] ) , we have to check that @xmath24 and @xmath25 remain finite when @xmath13 , which is not obvious at first glance . since @xmath51 by construction , a taylor expansion of @xmath52 around @xmath18 yields : @xmath53

there is apparently a widespread belief that the gravitational field ( and subsequently the rotation curve ) `` inside '' razor - thin , axially symmetric disks can not be determined accurately from elliptic integrals because of the singular kernel in the poisson integral . here , we report a simple and powerful method to achieve this task numerically using the technique of `` density splitting '' .

the evolution of baryons within their dark matter halos and the morphologies of the resulting systems depend on the merger and accretion history of the parent halos . major efforts of galaxy evolution studies today focus on how galaxies acquire the gas which fuels their star formation and what processes drive the distinctions between the red sequence and the blue cloud . still , our view of the extragalactic universe is only as complete as our methods for cataloging the galaxies that populate it . while the public wide area optical / ir and associated spectroscopic surveys are good at detecting luminous ellipticals , bright spirals and bursting or active galaxies , they are substantially less complete in tracing the low surface brightness , dwarf and gas - rich galaxy populations that actually dominate the local population . each catalog derived from an individual survey has its own built - in limitations and biases which affect our ability to construct a true census of the present day universe . because of its relatively simple physics , the hi line provides a useful tracer of the cool gas mass and of the star formation _ potential _ in nearby galaxies and probes the very population of modest luminosity , gas rich objects which are often underrepresented in surveys selected by optical / ir properties . while it is clear that most stars form out of molecular rather than atomic hydrogen , the molecular clouds themselves develop through the collapse of overdensities in the more diffuse , neutral medium thus , while the connection of hi to star formation is on small scales indirect , the global hi content serves as a tracer of relative sf potential . however , at present , hi line measurements yield hi masses @xmath14 for far fewer galaxies than those for which stellar masses @xmath15 are available from optical / ir wide area surveys . in fact , only now are hi surveys adequate in terms of volume sensitivity to sample a cosmologically significant volume @xcite . after the pioneering results delivered by small - scale surveys such as the arecibo hi strip survey ( ahiss : * ? ? ? * ) and the arecibo dual beam survey(adbs : * ? ? ? * ) , the advent of multi - feed array receivers on large single dish telescopes made possible wide - area 21 cm hi line surveys , such as the hi parkes all - sky survey , ( hipass : * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) and the companion hi jodrell bank all - sky survey ( hijass : * ? ? ? while covering a large fraction of the sky , these surveys failed to sample a cosmologically fair volume because their mean depth was too shallow , typically @xmath16 mpc , and they were limited in both angular and spectral resolution and in sensitivity . as a result , hipass sampled only sparsely both the most hi - rich but rare objects and the lowest halo mass systems detectable only if very nearby and with very narrow hi line widths and , because of the large parkes antenna beam ( 15.5 ) , suffered from confusion in the identification of optical counterparts ( ocs ) . the advent of a similar seven feed array at arecibo ( `` alfa '' , the arecibo l - band feed array ) has enabled a second - generation wide area extragalactic hi line survey , alfalfa , the * a*recibo * l*egacy * f*ast * alfa * survey @xcite . initiated in february 2005 , survey observations are now more than 90% complete . in this paper , we present the catalog of hi detections covering about 40% of the planned survey sky area , referred to hereafter as the @xmath2 catalog . both by design and because of improvements made possible by the accumulation and analysis of more survey data , the catalog presented here both extends and supercedes earlier ones presented by @xcite . in addition , the alfalfa data release presented here includes , where applicable , a cross reference to the optical survey dataset corresponding to data release 7 ( dr7 ) of the sloan digital sky survey ( sdss : * ? ? ? * ) . the availability now of a large body of alfalfa data , constituting 40% of the expected final survey , allows us to undertake an examination of the characteristics of its catalog of hi sources . @xcite and @xcite have presented earlier considerations of survey characteristics for subsets of the @xmath2 catalog specifically in the context of using the alfalfa survey to derive the hi mass function ( himf ) and to establish a standard of normal hi content for galaxies in low density environments , respectively . here , we examine the full @xmath2 catalog , discuss its identification of optical counterparts , and compare parameters derived from its measurements with those available in the previous compilation of targeted hi line observations presented by @xcite . we also present a more detailed look at the completeness of @xmath2 and how hi source catalog limitations in general can affect measurements of the himf . this paper is organized as follows : in [ sec : observations ] , we discuss the observational strategy , sky coverage , and data processing associated with the production of the alfalfa dataset and its final data products . [ sec : datapres ] presents the @xmath2 catalog of hi sources . the identification of the optical counterparts ( ocs ) of the hi sources is discussed in [ sec : oc ] . in that section , we present the crossmatch of the @xmath2 catalog to the sdss dr7 database and discuss those circumstances under which the alfalfa detection is not associated with an oc . a comparison of the hi line parameters derived from the alfalfa survey with those extracted from the large targeted hi dataset presented in @xcite is used in [ sec : validate ] to validate the photometric and spectral calibration underlying the alfalfa source parameters . an analysis of the survey completeness and reliability is presented in [ sec : completeness ] followed in [ sec : impact ] with a discussion of how the @xmath2 survey characteristics impact its cosmological applications , in particular , the derivation of the himf . a brief summary of the main points of this paper is given in [ sec : conclusion ] . the alfalfa observing strategy has been discussed in detail in @xcite and @xcite . of particular note to this data release , observations during a given observing session use the alfa seven - beam receiver parked on the meridian with data acquired in `` almost fixed '' drift - scan mode ; minor motion of the telescope is permitted so that the position of the central beam tracks in constant j2000 declination . with the feed arm positioned along the meridian at azimuths near 180@xmath17 ( for declinations north of the arecibo zenith at dec . = 18@xmath1721 ) or 360@xmath17 ( for declinations south of zenith ) , the feed array is rotated by 19@xmath17 so that the seven beams sweep out tracks equally spaced in declination by about 2.1 . in nearly all circumstances , a given observing run is dedicated to a single declination track . the 2-d ( time versus frequency ) drift scan datasets are converted from fits to idl format and run through an initial bandpass calibration and subtraction , normally within 24 hours of acquisition . in contrast to traditional total power , position switched pointed observations , a drift - scan survey ( of which alfalfa is certainly not the first example ) collects spectra continuously ( almost ) without moving the telescope . in the case of the alfalfa survey , the sampling rate is 1 hz , i.e. a spectrum of 4096 spectral channels ( a `` record '' ) is recorded every second for each polarization of every beam of the feed array . the slowly - changing characteristics of the bandpass with time can thus be monitored effectively . the alfalfa pipeline does so by separately monitoring the behavior of each spectral channel across the time domain , through a robust , low - order polynomial fit ( which skips over sources ) , outside of the spectral region dominated by galactic emission . for each 600 record unit ( a 10 minute drift `` scan '' ) , we thus obtain a two - dimensional map of the bandpass which can be `` subtracted '' from each spectral record . such `` sky subtraction '' is thus conceptually similar to that of the traditional position - switching mode , although the duration of the `` off '' is much larger than that `` on '' source , gaining @xmath18 in sensitivity with respect to standard position - switching observations . during the same processing step , continuum subtraction is also performed , and a separate continuum map is recorded . for spectral channels affected by galactic hi emission , such `` sky subtraction '' is not an option , and the bandpass subtraction can not be applied in the same manner as for spectral channels away from the galactic signal . in this case , the spectral shape of the bandpass across the galactic emission region is adopted as a linear interpolation between the two galactic emission - free sides of the spectrum . thus , the flux calibration of galactic features processed by the standard alfalfa pipeline is not accurate . each 2-d bandpass - subtracted dataset for each beam and each polarization is examined interactively and flagged for radio frequency interference ( rfi ) ; regions characterized by lowered quality ( due to standing waves , gain instabilities etc ) are assigned a lower weight . while this step ( known as _ `` flagbb '' _ ) is laborious , the facts that the continuum information is retained and the rfi is not median filtered away enables the further use of the dataset to look for hi absorption , for the derivation of upper limits at arbitrary positions in 3-d , and for stacking analysis @xcite . the flattened and flagged 2-d line and continuum maps are archived as level i datasets . once the set of drift scans providing full coverage for a complete strip in declination is flagged in this manner , the set of evenly gridded data cubes is generated . details of the gridding process are given in @xcite and summarized here . the grids are square in the angular dimension , 2.4@xmath17 on a side , evenly sampled at 1 spacing . their center positions on the sky are spaced 8@xmath19 apart in r.a . and centered on odd integer declinations ; the spatial dimensions of a grid are 144 by 144 pixels . for convenient access using modest data processors , each spatial grid is split into four , partially overlapping subgrids , each covering 1024 frequency channels . the grid generation algorithm also converts the spectral intensities from units of antenna temperature to mjy / beam in flux density , correcting for zenith angle variations in the gain of the telescope . a first step in the examination of the grids performs an astrometric fit to the continuum sources within them ; this fit is then used to subtract off the residual telescope pointing errors @xcite . grids are then flatfielded and rebaselined in both the angular and spectral dimensions to improve their quality by accounting for variations in gain , calibration and other systematic blemishes . `` flatfielding '' here corresponds to the process by which pixel - to - pixel variations within each channel map , caused mainly by continuum fluctuations , are accounted for . for spectral channels away from galactic emission , extragalactic hi sources are typically small in comparison with the angular size of alfalfa data cubes ( `` grids '' of @xmath20 ) . scale variations in the continuum level which may not have been effectively removed by the bandpass subtraction procedure can be identified by robust - fitting a two - dimensional surface ( in the angular domain ) from the channel map . in the absence of very strong continuum sources , this correction is generally small and it does not affect noise statistics in any significant way . after the angular flat fielding is performed , residual , localized spectral baseline features are also removed by subtracting low order polynomial fits to the signal free portions of the spectral domain around emission features . these arise , for example , from standing waves produced by multiple reflections of continuum source emission within the optical path . signal extraction is applied following @xcite , and once a catalog of candidate detections has been obtained , the grid is interactively examined , the global profiles are extracted , fluxes are measured , ocs are identified and remarks are recorded . it should be noted that this interactive process improves the definition of source parameters beyond the model fitting used by the automatic signal extractor ; this point , and the resultant reliability and completeness of the catalog , is discussed more fully in [ sec : impact ] . the final catalog of sources is constructed following a process of culling poorer quality detections where a source is contained in adjacent overlapping grids and running a series of data quality checks . the catalog presented here supercedes previous alfalfa data releases for several reasons mainly having to do with ( 1 ) the increased size of the available dataset which yields better understanding of pointing errors , gain variations and other instrumental artifacts , ( 2 ) improved sdss coverage since the first catalogs were produced , ( 3 ) improvements in the algorithm used to make global profile measurements and ( 4 ) increased contiguous coverage . some earlier measurements tended to underestimate fluxes for the brightest and more extended sources , a systematic effect for which a correction is now applied ( see [ sec : validate ] for the comparison of flux density measurements with published values ) . in most cases , changes to the flux density measurements included in earlier data releases are minor , but the current catalog is intended to replace the earlier ones entirely . it should be noted that further revisions of parameters for sources located near edges of the current grid coverage will come in the future in those cases when a newer grid in an adjacent strip better encompasses the source or contributes a higher quality dataset . by its nature as a cumulative drift scan survey , the harvest of alfalfa will both grow and improve over time . the full alfalfa survey is intended to cover 7000 deg@xmath1 of sky in two regions of high galactic latitude within @xmath21 of the arecibo zenith . all declinations will be covered @xmath22 dec . @xmath23 . since all observations are conducted during nighttime hours , the two regions are referred to as `` spring '' and fall. the `` spring '' region extends from @xmath3 r.a . @xmath4 while the `` fall '' alfalfa region encompasses from @xmath9 r.a . some sources are found outside the stated r.a . boundaries where the actual drift scan observations extended beyond the nominal map area . some priority has been given to completing areas within the sdss spectroscopic survey footprint , and the pace of observing has been dictated by the availability of telescope time . figure [ fig : aitoff ] illustrates the area of the sky contained in the @xmath2 catalog presented here : regions @xmath3 r.a . @xmath4 , @xmath5 dec . @xmath6 and @xmath7 dec . @xmath8 ( the `` spring '' region ) and @xmath9 r.a . @xmath10 , @xmath11 dec . @xmath6 and @xmath7 dec . @xmath12 ( the `` fall '' region ) . we present in table [ tab : catalog ] the measured parameters for 15855 detections , 15041 of which are certainly associated with extragalactic objects . an additional 814 are detected at velocities which suggest they may not be extragalactic but are more likely to be galactic high velocity cloud ( hvc ) features . the contents of table [ tab : catalog ] are as follows : * col . 1 : entry number in the arecibo general catalog ( agc ) , a private database of extragalactic objects maintained by m.p.h . and r.g . the agc entry normally corresponds both to the oc and the hi line source except in the cases of hvcs and other hi sources which can not be associated with an optical object with any high degree of probability . in those cases , the agc number corresponds only to the hi detection . an agc number is assigned to all alfalfa sources ; it is intended to be used as the basic cross reference for identifying and tracking alfalfa sources as new data acquired in overlapping regions supercedes older results . note that in previous alfalfa catalogs , an index number was used , a practice no longer employed ; a cross - reference to these older identifications is provided in table [ tab : comments ] . the designation of an alfalfa source referring only to its hi emission ( without regard to its oc ) should be given using the prefix `` hi '' followed by the position of the hi centroid as given in col . 3 of table [ tab : catalog ] . 2 : common name of the associated oc , where applicable . further discussion of the process of assigning optical counterparts is presented in [ sec : ocfind ] . * 3 : centroid ( j2000 ) of the hi line source , in hhmmss.ssddmmss , after correction for systematic telescope pointing errors , which are on the order of 20 and depend on declination . the systematic pointing corrections are derived from an astrometric solution for the nrao very large array sky survey ( nvss ) radio continuum sources @xcite found in the grids . as discussed in @xcite and @xcite , the assessment of centroiding errors is complicated by the nature of 3-d grid construction from the 2-d drift scans , those often acquired in widely separated observing runs , and , for resolved / confused sources , unknown source structure . as those authors suggest the best assessment of hi centroid error is accomplished by comparison of the hi centroids with the positions of the adopted ocs . an analysis of the positional offsets of the hi centroids from the positions of the ocs yields a relation for the median error in the hi position err@xmath24 as a function of the signal - to - noise ratio , s / n ( see col . 7 ) , for the @xmath2 sample : @xmath25 on average , the positional offset is about 18 , but it can , in rare instances exceed 1 ; those cases are noted in the comments included in table [ tab : comments ] . * 4 : centroid ( j2000 ) of the most probable oc , in hhmmss.ssddmmss , associated with the hi line source , where applicable . the oc has been identified and its likelihood has been assessed interactively using tools provided through the _ skyview _ website or the sdss explore tool , in addition to to the nasa extragalactic database ( ned ) and the agc and make use of judgmental criteria including redshift ( when known ) , size , morphology and optical color . the optical positions are normally estimated to be 3 or better but may be larger in exceptional cases ( very low surface brightness or peculiar , disturbed objects ) . the process of assignment of the most probable oc is discussed in [ sec : ocfind ] . it should be noted that only one oc is assigned per hi source although in reality confusion within the telescope beam is a possibility . suspected cases of confusion or ambigous assignment of the oc are noted in the comments included in table [ tab : comments ] . * col . 5 : heliocentric velocity of the hi source , @xmath26 in km s@xmath27 , measured as the midpoint between the channels at which the flux density drops to 50% of each of the two peaks ( or of one , if only one is present ) at each side of the spectral feature ; see also @xcite . the error on @xmath28 to be adopted is half the error on the width , tabulated in col . * col . 6 : velocity width of the hi line profile , @xmath29 in km s@xmath27 , measured at the 50% level of each of the two peaks , as described in col . 5 and corrected for instrumental broadening . no corrections due to turbulent motions , disk inclination or cosmological effects are applied . the estimated error on the velocity width , @xmath30 , in km s@xmath27 , follows , in parentheses . this error is the sum in quadrature of two components : a statistical error and a systematic error associated with the subjective guess with which the person performing parameter extraction estimates the spectral boundaries of the feature , flagged during the interactive assessment of candidate detections . in the majority of cases , the systematic error is significantly smaller than the statistical error ; thus the former is ignored . * col . 7 : integrated hi line flux density of the source , @xmath31 , in jy km s@xmath27 . this value corresponds to the total hi line flux measured on the integrated spectrum obtained by spatially integrating the source image over a solid angle of at least @xmath32@xmath33 and dividing by the sum of the survey beam values over the same set of image pixels ( see * ? ? ? * ; * ? ? ? estimates of integrated flux densities for very extended sources with significant angular asymmetries can be misestimated by our algorithm , which is optimized for measuring sources comparable with or smaller than the survey beam . a special catalog with parameters of extended sources will be produced after completion of the survey . the issue is especially severe for extended hvcs that exceed in size that of the alfalfa data cubes . in these specific cases , only the flux in the knots of emission is measured . in general , the hvcs have been catalogued here applying the same kind of s / n selection threshold as for the extragalactic signals , with the exception of the southern extension of wright s cloud , where , in addition to a bulk measurement of the portion of the cloud lying within this region , a selection of the brightest knots was measured to trace the structure . see column 12 and the corresponding comments for individual objects . the estimated uncertainty of the integrated flux density , in jy km s@xmath27 , is given in parentheses . * col . 8 : signal to noise ratio s / n of the detection , estimated as @xmath34 where @xmath31 is the integrated flux density in jy km s@xmath27 , as listed in col . 7 ; the ratio @xmath35 is the mean flux density across the feature in mjy ; @xmath36 is either @xmath37 for @xmath38 km s@xmath27 or @xmath39 for @xmath40 km s@xmath27(@xmath36 is a smoothing width expressed as the number of spectral resolution bins of 10 km s@xmath27 bridging half of the signal width ; the raw spectra are sampled at 24.4 khz @xmath0 5.5 km s@xmath27 at @xmath41 ) ; and @xmath42 is the r.m.s noise figure across the spectrum measured in mjy at 10 km s@xmath27 resolution , as tabulated in col . 9 . * col . 9 : noise figure of the spatially integrated spectral profile , @xmath42 , in mjy . the noise figure as tabulated is the r.m.s . as measured over the signal and rfi free portions of the spectrum , after hanning smoothing to a spectral resolution of 10 km s@xmath27 . * 10 : adopted distance in mpc , @xmath43 . for objects with @xmath44 km s@xmath27 , the distance is simply estimated as @xmath45 where @xmath46 is the recessional velocity measured in the cosmic microwave background reference frame @xcite and @xmath47 is the hubble constant , adopted to be 70 km s@xmath27mpc@xmath27 . for objects with @xmath48 km s@xmath27 , we use the local universe peculiar velocity model of @xcite , which is based on data from the sfi++ catalog of galaxies @xcite and results from analysis of the peculiar motions of galaxies , groups , and clusters , using a combination of primary distances from the literature and secondary distances from the tully - fisher relation . the resulting model includes two attractors , with infall onto the virgo cluster and the hydra - centaurus supercluster , as well as a quadrupole and a dipole component . the transition from one distance estimation method to the other is selected to be at @xmath49 km s@xmath27 because the uncertainties in each method become comparable at that distance . where available , primary distances as available in the published literature are adopted . when the galaxy is a known member of a group @xcite , the group systemic recessional velocity @xmath46 is used to determine the distance estimate according to the general prescription just described . * 11 : logarithm of the hi mass @xmath14 , in solar units , computed via the standard formula @xmath50 and assuming the distance given in col . 10 . no correction for hi self - absorption has been applied . * 12 : this column contains three relevant coded flags : + the first code , assigned as an integer value of 1 , 2 or 9 , refers to the category of the hi detection defined as follows : + code 1 refers to sources of s / n and general qualities that make it a reliable detection . these signals exhibit a good match between the two independent polarizations observed by alfalfa , a spatial extent consistent with the telescope beam ( or larger ) , an rfi - free spectral profile , and an approximate minimum s / n threshold of 6.5 @xcite . these criteria lead to the exclusion of some candidate detections with s / n @xmath51 ; likewise , some features with s / n slightly below this soft threshold are included , due to optimal overall characteristics of the feature , such as well - defined spatial extent , broad velocity width , and obvious association with an oc . we estimate that the detections with code 1 in table [ tab : catalog ] are nearly 100% reliable ; the completeness and reliability of the @xmath2 catalog are discussed in [ sec : impact ] . + code 2 refers to sources categorized as `` priors '' . they are sources of low s / n ( @xmath52 6.5 ) , which would ordinarily not be considered reliable detections by the criteria set for code 1 , but which have been matched with ocs with known optical redshifts coincident ( to within their errors ) with those measured in the hi line . we include them in our catalog because they are very likely to be real . in general , however , they should not be used in statistical studies which require well - defined completeness limits ; this point is further discussed in [ sec : impact ] . + code 9 refers to objects assumed to be hvcs ; no estimate of their distances is made . + of the 15855 sources included in this data release , 11941 are classified as source code 1 , 3100 are code 2 , and 814 are code 9 . + the second code , assigned as an alphabetic character , refers to a category reflecting the status of the cross identification of the alfalfa detection with an entry in the sdss dr7 database , as judged by the alfalfa team . this code is used to identify galaxies which lie outside the sdss dr7 sky footprint or for which there are clearly issues with the identification . it should be noted that this code refers only to the cross match with sdss dr7 . the cross - reference and basic parameters of the ocs is given in table [ tab : sdsscross ] . this code and its interpretation are as follows : + i : `` identified '' : the photoobjid is set but no other indicative flags have been applied ; this code applies whether or not there is a sdss spectroscopic counterpart . + o : `` outside dr7 '' : the sdss oc lies outside of the sdss dr7 footprint and thus no dr7 crossmatch can be performed . + u : `` unidentified '' : no sdss oc has been identified , but the object lies within the sdss dr7 footprint . + n : `` no dr7 photometric i d '' : no sdss dr7 photometric source has been identified ; assignment of this code can result from proximity to bright star , satellite trails , incomplete coverage or for other reasons . + m : `` missing '' : the oc is in the sdss dr7 footprint region but neither a photoobjid or a specobjid are returned to queries of the sdss dr7 database . + p : `` photometry suspect '' : the sdss dr7 photometry for the associated photoobjid are suspect for some reason as judged by the alfalfa team . assignment of this code often is associated with the identification of multiple near - equal - flux photometric objects within an obviously single oc . such cases apply often to very large optical objects or to faint , low surface brightness and/or patchy systems . the optical photometry associated with the sdss `` parent '' object may be adequate but caution should be exercised . + d : `` displaced sdss object '' : the sdss photo / spectid is displaced from the optical galaxy center , as identified by alfalfa team . the photoobjid may be legitimate ; often this is brightest photometric `` child '' . because of the displacement , the sdss redshift may not reflect the systemic recessional velocity of the galaxy . + t : `` two sdss objects '' : the sdss photoobjid associated with the galaxy center is displaced from the target associated with the sdss spectobjid , as judged by the alfalfa team , i.e. , the best photoobjid does not coincide with the spectobjid . usually , the spectobjid is an offcenter hii region or other bright knot within the target galaxy . + s : `` superposed sdss object '' : the sdss redshift corresponds to a superposed foreground star or background qso . + b : `` bad sdss solution '' the sdss redshift is unreliable or rejected for some unspecified reason . + the third code , given as an asterisk where applicable , indicates that a comment regarding the hi detection and/or the assignment of the oc is included for this source in table [ tab : comments ] . only the first few entries of table [ tab : catalog ] are listed in the printed version of this paper . the full content of table [ tab : catalog ] is accessible through the electronic version of the paper and will be made available also through our public digital archive site and the alfalfa project data site . in addition to the hi emission sources presented in table [ tab : catalog ] , it is expected that the alfalfa spectral cubes will also contain evidence for hi in absorption . @xcite discuss a pilot program which uses an adaptation of the alfalfa pipeline to search for hi absorption along the line of sight to nvss sources in a small number of the alfalfa cubes . the known hi absorber in the interacting system ugc 6081 was recovered . because the standard alfalfa reduction is not designed to look for such phenomena , the hi absorption detection is not included in table [ tab : catalog ] , and the reader is referred to @xcite for its parameters . table [ tab : comments ] contains comments about entries in table [ tab : catalog ] which have been recorded in the course of extracting source parameters and identifying the ocs . the second column contains a cross - reference to the catalog identification used in earlier papers , which is no longer used . we repeat in column 3 the hi detection code assigned for each source ( the first code in column 12 of table [ tab : catalog ] described above ) . it should be noted that angular separations given in these comments are referenced to the centroid of the hi source , not the position of the oc . these notes are somewhat heterogeneous in nature , having been incorporated during the process of data reduction by the individual responsible for source extraction . since the extraction has been performed over a period of several years , the databases available to the person making the comments have evolved ; thus the mention of nearby neighbors is not intended to be complete and should not be used in any derivation of local density . in some cases these notes identify issues with data quality , certainty of the oc or parameter extraction . the presence of a note does not mean necessarily that parameters are less certain than their errors indicate , as we have a tendency to err on the conservative side of casting doubt . they are included here because they provide an additional contribution to the legacy value of the dataset . subsets of the @xmath2 catalog have been included in the derivation of the hi mass function @xcite and the hi width function @xcite ; both papers include discussion of the sample characteristics , limitations and biases . similar to figures shown in those papers , figure [ fig : histograms ] illustrates the distributions of ( top to bottom ) redshift @xmath53 , @xmath29 , log @xmath31 , log s / n and log @xmath14 for the full alfalfa @xmath2 sample presented in table [ tab : catalog ] , while figure [ fig : spanplot ] shows the corresponding spaenhauer plot . further discussion of the impact of survey characteristics on cosmological issues and specifically on the derivation of the hi mass function is given in [ sec : impact ] . the principal aim of alfalfa is to catalog all gas - bearing extragalactic objects in the local universe . an integral part of understanding this hi census is similarly identifying the stellar counterpart associated with each hi source , or even more importantly , rejecting that such a counterpart exists . during the alfalfa data reduction process , optical images from the palomar digital sky survey ( dss2 ) and , where available , the sdss are interactively examined alongside the alfalfa hi dataset and the most probable oc is identified and recorded . while this assignment may not be correct in individual cases , it provides a first approach to understanding the relationship between the hi source and its stellar counterpart . the notes included in table [ tab : comments ] record comments on this process made by the alfalfa team member performing this interactive stage of the data analysis . in this section , we describe the process by which ocs are identified and discuss unresolved issues , provide a cross reference of sources to the sdss dr7 database and summarize general results on the evidence for `` optically - dark '' galaxies . we make use of virtual observatory tools embedded in the idl - based alfalfa reduction package @xcite to access several public imaging and catalog databases at several stages in the data reduction process . during the process of hi parameter measurement ( the routine called _ `` galflux '' _ ) , both dss2(b ) and sdss images are examined to identify interactively the most probable oc of each alfalfa source . because of their generally superior quality and ancillary information , preference is given to the sdss images where they are available . entries in our internal agc database as well as those listed in the nasa extragalactic database ( ned ) can be retrieved and examined . the alfalfa team member processing each source uses the available public information as to color , morphology , redshift , separation from the hi centroid in combination with his / her scientific judgement in assigning an optical counterpart . it should be noted however that because the catalogued data presented here were reduced over a three year time period , not all current information / data were available at the time this assignment was made . consistency checks are made later to look for redshift discrepancies or cases of large positional offset . with that caveat in mind , figure [ fig : ocfind ] shows several examples which illustrate the process of identification of ocs and the uncertainties inherent in it . each panel shows a 3 by 3 sdss g - band image centered on the hi centroid . the superposed circle marks the oc identified in table [ tab : catalog ] ; the size of each circle is arbitrarily chosen for best illustration of the target . the panels are intended to illustrate some of the challenges of assigning the oc by highlighting four specific cases as follows : * the upper left image is centered on the best - fit position of the hi source detected at hi095452.2,+142907 , a weak source of s / n = 7.3 . the corresponding oc agc 193821 is identified as the small galaxy sdss j095453.79 + 142910.0 22 from the hi centroid and partly contaminated by the diffraction spike of the bright foreground star ; the galaxy is more evident in the dss2(b ) image . there is no further optical information . * the upper right image is centered on the position of hi123120.9 + 050402 , a marginal alfalfa detection with a s / n = 4.9 . the oc agc 220720 is identified as vcc 1347 = cgcg 042 - 143 = j123117.00 + 050429.3 , a small spiral galaxy offset from the hi centroid by about 64 ; the large offset is not surprising given the low s / n of the hi detection . the sdss optical redshift is 9830 @xmath54 30 km s@xmath27 , slightly off the hi @xmath53 of 9873 @xmath54 4 km s@xmath27 . because of the low s / n of the hi emission profile but the coincidence with an optical galaxy with an adequately close redshift match , the optical identification is made and the source is designated as a `` prior '' and assigned an alfalfa detection category code of 2 . * the lower left image is centered on the position of hi 152240.3 + 055017 , a very narrow ( @xmath55 km s@xmath27 ) feature at @xmath53 = 1796 km s@xmath27 . the oc is identified as a dwarf galaxy agc 258471 better evident in the dss2(b ) image at j152238.7 + 054945 ; the sdss pipeline identifies at least five photometric objects within the lsb emission associated with the dwarf so that its magnitude is poorly measured . the offset of the hi centroid from the optical object is about 38 . * the lower right image is centered on the position of hi 160743.9 + 272201 , a source of s / n = 10.9 . as evident in the sdss g - band images , there are several objects in the field , including a close pair associated with sdss spectroscopic target j160744.75 + 272140.2 = kug 1605 + 275 ned01 with a redshift from the sdss of 23676 @xmath54 31 km s@xmath27 . the redshift is too high to be associated with the alfalfa hi source ; several other galaxies in the vicinity of this system have similar redshifts . careful examination of the sdss image shows a second object , which is not identified in the sdss photometric database and which appears to be partly overlapping with j160744.75 + 272140.2 but in its foreground at j160743.9 + 272201 . we identify the hi source with this foreground blue galaxy which becomes agc 749361 . we emphasize again that because of the alfalfa centroid position uncertainty and its relatively large beam size , assignment of the most probable oc is a reasonable but not a perfect process . furthermore , it will continue to be a dynamic one , striving for improvement when new data provide improved detail . for example , the current dataset does not include yet a systematic incorporation of data from the sdss iii survey or its dr8 . although not available at the time of earlier alfalfa data releases , the completion of the sdss legacy survey has afforded the opportunity to cross reference the alfalfa and sdss datasets where the two share footprints . as a new feature of this and future alfalfa catalog releases , here we provide , in table [ tab : sdsscross ] , the cross identifications of alfalfa sources with the photometric and spectroscopic catalogs associated with the sdss , in this instance , with the data release dr7 @xcite . entries in table [ tab : sdsscross ] are as follows : * col . 1 : the source agc number , identical to col . 1 of table [ tab : catalog ] * col . 2 : the hi detection category code , identical to the 1st ( integer ) code in col . 12 of table [ tab : catalog ] * col . 3 : the sdss cross reference catagory , identical to the 2nd code in col . 12 of table [ tab : catalog ] * col . 4 : the sdss dr7 photometric catalog object identification number ( photoobjid ) , where applicable . * col . 5 : the sdss dr7 spectroscopic catalog object identification number ( specobjid ) , where applicable . * col . 6 : the r - band model magnitude corresponding to the photometric object or its sdss parent . * col . 7 : the ( u - r ) color associated with the oc from the sdss as reported in the dr7 . this value is used in figure [ fig : baldryplot ] , and in order to allow direct comparison with figure 9 of @xcite , it has not been corrected for extinction or redshift . * 8 : the redshift corresponding to the sdss dr7 spectroscopic catalog object , extracted from the sdss dr7 database , where applicable . 9 : the error on the redshift given in col . 8 , extracted from the sdss dr7 database , where applicable . it is important that potential users understand the limitations associated with this alfalfa sdss cross - reference . as noted by @xcite and discussed in [ sec : datapres ] , the alfalfa hi centroid accuracy is of order 20 , but increases as the s / n decreases , as given in equation [ eq : poserr ] . furthermore , as is well known , the standard sdss image reduction pipeline suffers from source blending , and more importantly , shredding , particularly in the sources whose light distributions are patchy or of low surface brightness . the current alfalfa reduction process includes an interactive step of direct examination of the sdss imagery and issues associated with blending / shredding are noted immediately . however , earlier alfalfa datasets which predated release of dr7 were not subject to such individual cross examination . while attempts have been made to flag and check suspicious cases , it is likely that some misidentifications remain . the dr7 photometric catalog object identification number given in column ( 4 ) is the photoobjid whose magnitude and position given in sdss dr7 corresponds most likely to the oc ; the actual `` best magnitude '' may be associated with the sdss pipeline `` parent '' . users are cautioned to understand fully issues associated with blending , shredding and poor sky subtraction and to make use of warning flags and other quality indicators when using the photometry associated with the photometric object given here . particularly relevant discussions of background subtraction issues are given in @xcite and @xcite . similarly , the spectroscopic identification refers to the most probable and most closely related sdss spectroscopic target . this cross match likewise can suffer from issues of positional offset , signal - to - noise etc . and should be treated with similar caution . the sdss dr7 cross reference category given in col . 3 . of table [ tab : sdsscross ] ( and also as one of the two codes given in col . 12 of table [ tab : catalog ] ) provides further comment on quality issues as identified by members of the alfalfa team . however , because some of the processing of alfalfa data predates the release of sdss dr7 , this code assignment should not be considered complete : many but not all sources have been revisited after the release of dr7 . the intent of providing the cross reference is to make statistical studies more convenient and potentially homogeneous . but again , we emphasize the importance of visual examination of individual cases where such attention is critical to the drawing of scientific conclusions . of the 15041 extragalactic ( i.e. , non - hvc ) objects listed in table [ tab : catalog ] , 2312 lie outside the sdss dr7 footprint and 199 are classified as `` dark '' ( see [ sec : darkgals ] ) . of the ones with identified ocs and included in dr7 , 11740 are assigned sdss code `` i '' ( meaning the sdss photometric identification is acceptable and there is no issue with the spectroscopic identification where such exists ) while the others are given a code in table [ tab : sdsscross ] indicating a recognized issue with either the sdss photometry or spectroscopy . the alfalfa fall portion of the sky contains some regions for which only photometry is available ; in the spring region , the photometric and spectroscopic footprints overlap more completely . of the 11240 alfalfa spring sky galaxies with a corresponding sdss photometric i d ( of any code ) , 9377 ( 83% ) have an associated entry in the sdss spectroscopic catalog , and 1863 ( 17% ) do not . figures [ fig : conespr26 ] and [ fig : conefall26 ] provide graphical illustrations of the relative strengths and weakness of the alfalfa and the sdss surveys as tracers of the large scale structure in the local universe . figure [ fig : conespr26 ] shows a cone diagram of a four degree wide slice of the alfalfa spring sky centered on dec . = + 26@xmath17 and including the full alfalfa bandpass redshift range @xmath56 18000 km s@xmath27 . the upper cone extends over the full @xmath53 range covered by alfalfa and the lower one , only the inner @xmath56 9000 km s@xmath27 . blue open circles mark the locations of galaxies detected by alfalfa , while red filled ones denote objects with redshifts from the sdss dr7 . the fall - off in the density of blue points follows the distribution seen in figure [ fig : spanplot ] . the `` finger of god '' radial line - up of optical@xmath53 ( red ) points so prominent in the lower diagram is the coma cluster abell 1656 . galaxies in that cluster are well known to be strongly hi deficient @xcite so that alfalfa detects very few of them . as indicated by the numbers superposed on the diagram , the number of sdss spectroscopic targets in the full alfalfa volume is about three times the number of alfalfa hi sources ; in the inner volume illustrated in the bottom diagram , that ratio drops to two and to @xmath57 for field galaxies at @xmath56 5000 km s@xmath27 . while strong bias against finding hi sources in the regions of rich clusters is clearly evident , the hi - bearing galaxies trace well the large - scale supercluster structures and include some of the most isolated objects found in this nearby volume . for comparison , figure [ fig : conefall26 ] shows a similar cone plot covering a four degree wide slice of the alfalfa fall sky centered on dec . = + 26@xmath17 . the sdss spectroscopic survey did not cover this region ; the red filled circles mark objects with optical redshifts available from the literature . in this part of the sky , alfalfa sources contribute the majority of redshifts even at its outer boundary . it should be noted that the slice of the sky sampled in figure [ fig : conespr26 ] covers a strongly overdense region of the local universe , the coma - abell 1367 supercluster , whereas the fall region lies to the south of the main filament of the pisces - perseus supercluster and includes a portion of the void in front of it . as in all studies of the local universe , the actual large scale structure contained in the survey volume can leave a strong imprint on the observed distribution of galaxies and their properties in limited samples . further discussion of the impact of large scale structure on cosmological inference is included in [ sec : impact ] . making use of the sdss cross reference tabulation , figure [ fig : baldryplot ] presents a color magnitude diagram ( cmd ) for the @xmath2sdss overlap sample for comparison with similar diagrams extracted from the sdss photometric survey alone . a similar cmd was independently constructed by @xcite for a sample of alfalfa galaxies found in low density environments . in figure [ fig : baldryplot ] , grayscale and contours show the distribution of the hi - selected sample and the axes correspond to the range illustrated in figure 9 of @xcite . the superposed dashed line shows the optimum divider used by @xcite to separate galaxies on the red sequence ( above the curve ) from those in the blue cloud ( below it ) and given by their equation 11 . for the purpose of comparison with their figure 9 , no corrections for redshift or internal extinction have been applied to the magnitudes used to construct figure [ fig : baldryplot ] . figure [ fig : baldryplot ] can also be compared with figure 4 of @xcite who used a large sample of galaxies from sdss dr7 and did apply a k - correction ; in their figure , those authors also categorize separately elliptical and spiral galaxies according to the sdss catalog parameter @xmath58 , the fraction of the galaxy s luminosity contributed by the devaucouleurs profile . clearly , the @xmath2 catalog is dominated by blue spiral galaxies and is strongly biased against the red sequence . as discussed by @xcite , some of the luminous , red objects are truly red , luminous and gas - bearing objects ; other luminous objects appear red because they are edge - on disks for which the internal extinction correction is significant . still , figure [ fig : baldryplot ] confirms the conclusion of @xcite that the most luminous gas - rich population includes a significant fraction of red galaxies . further discussion of the stellar and star forming properties as derived from sed fitting the photometry provided by the sdss in the optical and the fuv / nuv by the galaxy evolution explorer ( galex ) satellite for the alfalfa sample will be presented in @xcite and @xcite . one of the scientific drivers behind blind hi surveys is the possibility of contributing gas rich but optically `` dark '' galaxies to the extragalactic census . previous analyses by e.g. , @xcite , of the statistics of targeted hi line surveys have shown that such objects must be rare ; otherwise there would have been more sources detected serendipitously in the random off - source positions observed by the total power position - switching observing mode used for most of those earlier surveys . indeed , perhaps the best example of an optically dark galaxy is the southwest component of hi 1225 + 01 @xcite , but it is not a purely isolated object , being located on the outskirts of the virgo cluster and part of a binary system with its dwarf galaxy companion to the northeast . of the 4315 hi sources reported in the hipass catalog , 84% were identified with one or more possible ocs @xcite . most of the remainder are located at low galactic latitude where galactic extinction strongly inhibits the hunt for the stellar counterpart . in fact , @xcite investigated through followup observations the 13 hipass without ocs and with a@xmath59 1 mag and concluded that not a single one could be claimed as an isolated dark galaxy . some might be intergalactic in the sense of being associated with tidal debris fields or fragments of very extended hi disks , but always there were nearby , visible ( stellar ) objects at the same redshift . because of alfa s superior angular resolution at l - band in comparison with that of the parkes telescope ( 4 vs 15.5 ) , we are able to centroid the position of the alfalfa hi sources to better than 20 on average and to identify their ocs likewise with better surety . only 1013 of the 15855sources presented in table [ tab : catalog ] do not have assigned ocs . of those , 814 blank field objects have observed velocities which fall within the range characteristic of emission associated with some milky way population . all of these are assigned an hi source category code of 9 in column 12 of table [ tab : catalog ] . they are likely to be hvcs , although a few isolated objects with narrow velocity widths and small angular sizes ( either barely resolved or unresolved ) are candidate low mass extragalactic halos @xcite . their distribution and nature will be discussed elsewhere . of the remaining 199 hi sources ( @xmath60% of the total extragalactic population ) whose velocities suggest that they are truly extragalactic , we have individually examined closely the sdss and/or dss2(b ) fields to look for ocs ; comments derived from that examination are included in table [ tab : comments ] . some of these objects do not lie in the region covered by sdss , making the identification of ocs more difficult , but by design , only a few lie in regions of significant optical extinction . roughly 3/4 of the `` dark '' hi sources are located in fields where objects of similar redshift are found , albeit beyond the reasonable limits of coincidence given by the alfalfa pointing accuracy . a number can be linked to previously known extended hi distributions such as the leo ring @xcite , the tail of ngc 4254 @xcite , the extended tail of ngc 4532/ddo 137 @xcite or the intergroup gas found in the ngc 7448/7463/7464/7465 group @xcite . among the blank field hi detections with sdss data ( including dr8 ) and not contaminated by the presence of bright foreground stars , only about 50 remain as candidates to be isolated `` dark '' objects . these objects are the targets of a followup program that will confirm their reality as hi sources with the arecibo single pixel l - band receiver , localize the hi emission via hi synthesis observations and search for associated low surface brightness stellar emission via optical imaging . oh megamasers ( ohms ) are powerful line sources associated with the starburst nuclei in merging galaxy systems . @xcite has pointed out that oh megamasers at @xmath61 may contaminate a blind extragalactic survey such as alfalfa . an extremely rare phenomenon in the local universe , about 100 ohms are known out to a redshift of 0.265 @xcite . the main 18 cm oh lines occur at rest frequencies of 1665 and 1667 mhz respectively . in ohms , the emission at 1667.359 mhz dominates ; that line is redshifted in the alfalfa observing band for sources with @xmath62 . using the large targeted survey for ohms by @xcite as a baseline for the expected flux density and spectral characteristics of ohms , it is probable that a few of the alfalfa sources without ocs may in fact be ohms with @xmath62 . confirmation that an alfalfa source is in fact an appropriately redshifted ohm and not an optically dark hi galaxy will require follow - up hi synthesis observations to localize the line emitting region and optical / ir spectroscopy to confirm the redshift . already , however , there are four ohm candidates which can be identified as such because the line emission occurs at frequencies higher than 1422 mhz . hence , under the assumption that the emission arises from the hi 21 cm line , the observed @xmath53 is too largely blueshifted for plausible interpretation as an extragalactic or galactic hi source . the properties of these four objects are given in table [ tab : ohmtab ] and optical images obtained from either the sdss or dss2(b ) are shown in figure [ fig : ohms ] . the entries in table [ tab : ohmtab ] are as follows : * col 1 : entry number in the agc * col 2 : centroid ( j2000 ) of the emission line source , in hhmmss.ssddmmss , as in col . 3 of table [ tab : catalog ] . the designation of the candidate then adopts the identifier `` ohmcand '' plus this centroid position . * col 3 : position of the identified optical counterpart , in hhmmss.ssddmmss . * col 4 : @xmath63 , redshift of the optical counterpart , where known . * col 5 : @xmath64 , redshift of the candidate ohm assuming its emission is dominated by the oh line at 1667.359 mhz . * col 6 : @xmath65 , heliocentric velocity if the emission were associated with the hi line , in km s@xmath27 . * col 7 : @xmath66 , oh line flux density , in jy km s@xmath27 . * col 8 : s / n of the oh line emission , defined as in col . 8 of table [ tab : catalog ] . * col 9 : rms noise in the vicinity of the line emission , defined as in col . 9 of table [ tab : catalog ] . in all four instances , there is a small object visible in public imaging databases which can be identified as the likely optical counterpart : * agc 102708 = ohmcand000337.0 + 253215 is likely associated with sdss j000336.02 + 253204.0 , a very tiny object also evident in dss2(b ) . there is no ned entry or redshift measurement . * agc 102850 = ohmcand002958.8 + 305739 is likely associated with 2masx j00295817 + 3058322 , a well - formed spiral galaxy . there is no confirming redshift measurement . * agc 181310 = ohmcand082311.7 + 275157 is likely associated with sdss j082312.61 + 275139.8 , also known as iras 08201 + 2801 and 5c 07.206 , a known ulirg . for this single object , the optical redshift @xmath53 = 50314 km s@xmath27 , @xmath67 = 0.167830 @xmath54 0.000041 from the sdss confirms the identification as an ohm ; its ohm emission was previously discovered by @xcite . * agc 228040 = ohmcand124540.5 + 070337 is likely associated with sdss j124545.66 + 070347.3 , a spiral galaxy viewed at high inclination as evident in figure [ fig : ohms ] . no confirming redshift measurement is available . ohms may also be identified in the subset of low s / n sources not included in the current catalog because they do not meet the criteria of codes 1 and 2 . by the simplest argument based on the fraction of the usable alfalfa bandwidth above 1422 mhz and assuming that these four candidates are , in fact , ohms , it is possible that half of the `` dark galaxy '' candidates discussed in [ sec : darkgals ] might be ohms at 0.175 @xmath13 0.245 . a similar estimate comes from considering the @xmath2 volume and the ohm luminosity function at low z @xcite . a more systematic approach to the identification of ohms throughout the full bandpass and using the 3-d alfalfa dataset is currently being undertaken by members of the alfalfa collaboration . most targeted extragalactic hi line flux densities are extracted from spectra conducted using a total power position switching technique . as outlined in [ sec : observations ] , the alfalfa dataset is generated using a very different approach whereby alfa drift scan data are obtained months and sometimes years apart and without doppler tracking . the 2-d datasets ( frequency versus time ) for the two individual polarizations of each of the seven beams are bandpass subtracted and flagged for rfi . after the acquisition of all the drifts for a region of the sky , the 3-d spectra grid is then generated . as with any new survey , it is critical to verify that the spectral scales ( velocity and flux density ) at each grid point are accurate . the cornell digital hi archive presents a large compilation of digital hi line spectra obtained using pointed observations of optically - selected targets @xcite which have been digitally analyzed using similar algorithms to those adopted for alfalfa . because those spectra were obtained with a variety of single - dish telescopes and spectrometers , careful attention was paid to correct for instrumental effects such as pointing errors , source extent , instrumental broadening and spectral smoothing . corrections for the various effects were modeled and tested to produce a homogeneous catalog of extracted properties with their associated error estimates . here , we present the validation of the alfalfa velocity , velocity width and flux density scale by comparison of @xmath2 catalog parameters with the previous targeted hi line observations of sources which have been re - detected by alfalfa . of the 2073 galaxies which are contained both in the @xcite and the @xmath2 catalogs , 1887 are classified as alfalfa code 1 sources and 186 are code 2 detections . the alfalfa `` minimum - intrusion '' observing mode acquires data without doppler tracking , i.e. , in topocentric mode . heliocentric corrections are applied in the fourier domain , whereby the appropriate velocity shift at each point ( each spectrum associated with each one second record for each polarization of each beam ) is calculated , converted to a phase gradient across the bandpass and applied to the fourier transform of each spectrum . the inverse fourier transform then gives the spectrum in the heliocentric rest frame which is used thereafter to yield the systemic velocity @xmath53 and the hi profile velocity width @xmath29 . it is important to note that the specific definitions of hi systemic velocity and the global profile velocity width are not uniformly adopted in the literature . for alfalfa , we adopt the same convention as that used by @xcite , that is , polynomials are fit to each side of the two - horned profile and then @xmath53 and @xmath29 are measured at the level of 50% of the peak intensity on either horn : @xmath53 is then the midpoint and @xmath29 the full width at that level . where appropriate ( face - on galaxies ; dwarf systems ) , a single gaussian provides the best fit and is similarly measured . figure [ fig : velcompare ] illustrates the comparison of the two parameters for the @xmath2-@xcite hi archive overlap sample . in both panels , the vertical axis shows the residual alfalfa - hi archive . the occurrence of outliers is expected because ( a ) alfalfa spectra correspond only to 40 sec per beam of integration time on source , whereas the targeted spectra are generally of much longer integration ; ( b ) targeted spectra are affected by pointing errors either in the coordinates used to position the telescope or intrinsic telescope pointing inaccuracy ; and ( c ) blends with close companions where the pointed spectra were taken with smaller single dish telescopes . a few cases of gross disagreement are explained by errors in the velocity scales of very old hi data which were acquired in the days before significant information was written into data headers , when the setup of the backend electronics and spectrometer required physical cabling and hand dial - setting at the start of each observing run and when records of frequency offsets for different quadrants of the spectrometer were kept only on hand - written index cards ; these cases are noted in the comments in table [ tab : comments ] . the appearance in the lower panel of some outliers at relatively low @xmath29 reminds us that at low s / n or in the presence of residual baseline structure , broad widths may be underestimated . the dependence of the sensitivity on line width is discussed in [ sec : completeness ] . as evident in figure [ fig : velcompare ] , the comparison of the velocity scales reveals no systematic offsets and agreement within the expected errors . as discussed in @xcite , practical limitations and instrumental uncertainties restrict the accuracy with which hi line flux densities can be measured to not better than a few percent . despite regular calibration via the injection of a noise diode , drifts in the electronic gain , amplifier instabilities , sidelobe variations and standing waves ( caused by multiple reflections within the optical path of cosmic continuum sources or terrestrial rfi ) induce variations in the total power , while baseline irregularities and data loss due to rfi impact the measurement of flux density in noisy data . as discussed in @xcite , hi line flux densities derived from targeted ( pointed ) observations are typically accurate to not better than 15% ; older datasets taken when amplifiers were substantially less stable than today are probably accurate to not better than 25% . calibration of the alfalfa dataset is performed in two separate stages . first , during the course of an observing run , a noise diode , calibrated by the engineering staff in the lab , is fired once every 600 seconds . the data stream then includes a record with this additional power source ( the `` cal - on '' record ) . all observing runs contain at least 9 such calibrations ; longer ones may contain as many as 60 . a polynomial fit is performed to the ratio of the total power with the calibration diode on , versus when it is off , for the whole set for an observing block . this polynomial is then used to correct the individual records of the drift scan data . the second method of calibration is performed on the data after grid construction , making use of the radio continuum sources which they contain . a comparison is made of the flux densities of the source contained in the grids with published values in the nvss @xcite , and then an average correction factor is applied to tie the alfalfa flux scale to the nvss . further details on calibration of the alfalfa dataset are given in @xcite . even when gain corrections for frequency dependence and other effects are correctly calibrated out , hi line flux densities observed with single point observations must be corrected for beam dilution and , often , pointing errors . in addition to the inaccuracy of telescope pointing , particularly important in early arecibo observations , the input positions used to point the telescope were accurate only at the level of 0.51 level for some of the oldest observations used to acquire the archival data reanalyzed by @xcite , @xcite report both raw ( as observed ) and corrected values of the hi line flux density for galaxies observed via pointed observations of optically - selected targets . the true hi line flux density was derived by applying corrections for telescope pointing errors , errors in the positions used to point the telescope ( both of which apply more importantly to older datasets ) and for partial resolution by the telescope beam . the latter is derived by adopting a hybrid correction for source extent that is based on a modeled hi distribution scaled by the optical size and an average telescope beam power pattern . as a drift scan mapping survey , alfalfa flux densities are not subject to such corrections . figure [ fig : fluxes ] shows the comparison of the hi line flux densities measured by alfalfa with the values given in the @xcite hi archive . the latter are corrected for pointing and position errors and for source extent ( but not for internal hi absorption ) . the vertical axis shows the ratio of the hi line flux densities reported in the two catalogs . alfalfa detection code 1 objects are shown as blue open circles ; the lower s / n code 2 objects are shown in red filled circle . since the error in the hi line flux density for both surveys depends on the hi line flux density itself as well as the s / n of the spectrum and the magnitude of the corrections applied to the pointed data , the increasing scatter seen in the ratio at low hi line flux densities is as expected . when corrections for source extent are applied to the pointed data , the flux density scales are coincident within the errors . among the sources with highest hi line flux density , figure [ fig : fluxes ] suggests that the alfalfa flux algorithm may be missing some flux . the total flux should be recovered since alfalfa is a mapping survey but the hi line flux from very extended sources , especially those located towards the edges of the constructed grids could be lost due to the finite grid size and the bandpass subtraction and grid flattening schemes . for those , alternative processing tools from the standard pipeline will be developed , after completion of the main survey . in order to assess the contribution of very diffuse , extended hi in vicinity of nearby , isolated galaxies , @xcite and @xcite observed a carefully- selected sample of @xmath0100 galaxies with both the former 42 m and the green bank telescopes . as they note , the uncertainty in the hi line flux density for their high signal - to - noise data is @xmath68 1% ; on the other hand , the uncertainties contributed by fitting the polynomial baseline and defining the boundaries of the emission profile are considerably larger . because an unblocked aperture should deliver reduced standing waves and minimal stray radiation , flux densities measured with the gbt should be more accurate than ones measured with a complex instrument like arecibo . at this point , there are only 12 galaxies in common with the sample observed very accurately with the former 42 m telescope by haynes et al . ( 1998 ) , not enough for conclusive results . these issues will be explored in a future work . although it might be expected to serve as the better dataset to use in examining systematic uncertainty and testing for missing flux from extended / bright sources , the northern hipass survey @xcite does not in practice provide an adequate comparison sample for several reasons . first , as mentioned above , flux calibration uncertainties do not dominate most hi flux density errors ; baseline uncertainties , noise and beam effects do . a drawback of the northern hipass catalog is that some of it suffers from residual baseline ripple , particularly when observations were made during the daytime @xcite . secondly , the sensitivity difference between arecibo and parkes means that the s / n of most alfalfa/@xcite detections is typically much higher than that of hipass . this fact alone , on top of the baseline issues , give the @xcite spectra a significant advantage over hipass in terms of parameter accuracy . thirdly , although there are 1000 galaxies in the northern hipass dataset @xcite at dec . @xmath69 , only @xmath0350 lie in the overlap with the present @xmath2 catalog . lastly , there are no code 2 sources detected by hipass ( which is not surprising , given its much poorer sensitivity ) so we can not make the comparison between codes 1 and 2 . however , for the record , we include in the lower panel of figure [ fig : fluxes ] the similar comparison of flux densities from alfalfa and hipass ; clear cases of confusion within the parkes beam are not included in this analysis . curiously , alfalfa detects _ more _ flux density than hipass in some cases ; examination of those reveals that they are mainly alfalfa detections with broad @xmath29 for which hipass detects , at much lower s / n , a lower hi line flux density and a narrower @xmath29 , clearly missing some of the hi line emission detected by alfalfa . the practical exploitation of any survey requires an understanding of its source sensitivity , completeness and reliability . in comparison with previous blind hi surveys , alfalfa offers a much richer dataset which itself can be used to probe the robustness of its source catalog . source extraction and parameter measurement for alfalfa is performed in a two - step process , which includes automated as well as interactive procedures . initial source extraction is performed by a fully automatic matched - filtering method @xcite . the algorithm uses templates which vary in shape as a function of profile width ( gaussian for narrow profiles , hermite functions for broad profiles ) , and outperforms algorithms based on smoothing followed by peakfinding ( see figure 3 in @xcite ) . the reliability ( i.e. the fraction of detections that correspond to real sources ) of this automated method is estimated to be @xmath70% for sources with @xmath71 , a value that was determined by performing source extraction on regions of the alfalfa datacubes expected to be devoid of cosmic signals ( corresponding to the velocity range @xmath72 km s@xmath27 , see 5.4 and figure 8 in @xcite ) . source candidates are then visually inspected and source parameters are interactively measured and catalogued . it should be noted that the parameters of the final catalogued sources ( e.g. @xmath31 , @xmath73 , @xmath29 , etc . ) do not generally coincide with the values determined by the automatic signal extractor , because the two procedures use different definitions and calculation methodologies for the parameters @xcite , and the human intervention is designed to optimize the measurement accuracy and further improve the reliability of the catalog by rejecting spurious detections that correspond to low - level rfi , poorly sampled data and residual baseline fluctuations . we therefore expect the final reliability of alfalfa code 1 detections to be very close to 100% . however , as discussed by @xcite , the reliability of alfalfa sources extracted by the matched - filtering algorithm drops precipitously below a s / n of 6.5 . the code 2 sources included in table [ tab : catalog ] fall below this nominal alfalfa s / n detection threshold , but are included in the catalog because they coincide with an optical galaxy of known ( prior ) and coincident redshift . although these sources should not be included in statistical studies which require careful consideration of survey completeness and sensitivity limits , the vast majority are likely real hi line sources , and the gas in them will also contribute to the overall hi density in the local universe . hence , we include them in the following analysis of alfalfa completeness and their impact on measurements of cosmological parameters . the two step process used to identify , extract and measure the alfalfa detections presented in table [ tab : catalog ] results in a catalog of reliable detections that is dependent on _ both _ the integrated hi line flux density of a given source and its global hi line profile width , w@xmath74 . like all fixed integration - time spectroscopic surveys , alfalfa is more sensitive to narrow hi profiles than to broader ones of the same integrated line flux . based on the demonstrated performance during the single - pass precursor observations of the observing equipment and the signal extraction pipeline , @xcite predicted the specific relationship between the integrated flux density detection threshold ( s@xmath75 in jy km s@xmath27 ) and the profile width ( w@xmath74 , in km s@xmath27 ) of a source in terms of the s / n required for inclusion in the catalog : @xmath76 note that the above expression differs from that given in equation 2 of @xcite ( numerical factor of 0.15 vs. 0.22 ) because that work adopts the rms appropriate to the single - drift maps used in the precursor observations . one of the principal conclusions of @xcite was that the two - pass strategy adopted for alfalfa would improve on that employed by the precursor program by a factor of 1.5 . `` sensitivity '' is a qualitative term that can be defined in terms of the survey `` completeness '' . we refer to the completeness of the alfalfa survey as that fraction of cosmic sources of a given integrated flux density and within the survey solid angle that are detected by alfalfa and included in the @xmath77.40 catalog . other blind hi surveys ( e.g. adbs , hipass ) have estimated the completeness of their catalogs as a function of profile width ( @xmath29 ) by examining their ability to recover synthetic sources of known characteristics ( peak flux , @xmath73 , @xmath29 , etc . ) injected into the spectral cubes . one of the motivations for such an approach is to assess the reliability of sources in the presence of non - gaussian noise . as noted by @xcite ( see 5.4 and figure 8 of that paper ) , the impact of non - gaussian noise on the automatic signal extractor developed for alfalfa is generally minimal above s / n = 6.5 . its presence , principally in the form of the very broad spectral standing waves resulting from reflections in the telescope focal structure ( e.g. , * ? ? ? * ) , is responsible for the upturn at large @xmath29 in equation [ eq : eqnthresh ] . at the narrower widths , there is no evidence that a gaussian assumption is unfair . now that a significant alfalfa dataset exists , the data themselves can be used to derive the true sensitivity limits . the analysis of the real survey data is motivated both by a desire to use the real observables rather than predictions of the performance of the observing equipment and signal extraction pipeline , and especially by the fact that the alfalfa survey has actually outperformed its predictions , as discussed in appendix a of @xcite . hence , we follow a different method to determine the alfalfa completeness that makes no use of `` fake sources '' , but relies instead on the @xmath77.40 catalog itself . for a flux - limited sample from a uniformly distributed population , number counts will follow a power - law with an exponent of @xmath78 . we then can determine the onset of incompleteness when our data deviates from this form . briefly , the details of this method consist of the following steps : 1 . the code 1 sources are divided into 32 equally spaced bins in @xmath79 . 2 . for each width bin , we count the number n of detected sources in logarithmic intervals of flux density to determine the @xmath80 histogram ; apart from the impact of large scale structure in the survey volume , number counts are expected to follow a power - law with an exponent of @xmath78 . 3 . for each bin in @xmath79 , we plot @xmath81 versus @xmath31 ; see figure [ fig : lognlogs ] for three representative width bins . this distribution should be flat if all sources are accounted for . a downturn at low @xmath31 thus marks the onset of incompleteness . we fit an error function to each histogram ( red dashed lines ) and assume completeness over the well sampled range of @xmath31 over which the distribution shows a flat plateau . we calculate the integrated flux density where the alfalfa completeness crosses @xmath82% , @xmath83% , and @xmath84% ( vertical red lines mark the @xmath82% completeness in each bin ) . in practice , the distributions drop off in the same in the same way , such that the 50% and 25% limits occur at a constant offset in @xmath85 from the 90% value across all bins . the values of @xmath86 for each w@xmath74 bin are then fit with the combinaton of two straight lines , similar to equation [ eq : eqnthresh90c1 ] , with a break at @xmath87 km s@xmath27 . the resulting @xmath82% completeness limit ( red solid line in upper panel of figure [ fig : fluxvsw ] ) for code 1 sources can be expressed as : @xmath88 where @xmath31 is in jy km s@xmath27 and @xmath29 is in km s@xmath27 . as mentioned before , the 50% and 25% completeness limits occur at a constant offset from the 90% value . the derived offsets for the code 1 sources only are : @xmath89 of the 15041 extragalactic objects in the @xmath2 sample , 3100 are categorized as code 2 detections ( low signal - to - noise detections with prior optical detection ) . the lower panel of figure [ fig : fluxvsw ] shows the corresponding plot of the distribution of sources in the @xmath90 plane for the @xmath2 extragalactic catalog , including the code 2 detections which are shown as green symbols . these additional hi sources are expected to have a lower detection threshold , clearly evident in the lower panel of figure [ fig : fluxvsw ] . an analysis identical to the above can be performed including both the code 1 and 2 sources , yielding a relation for the combined catalog ( red solid , dash - dotted and dotted lines in the bottom panel of figure [ fig : fluxvsw ] ) : @xmath91 and @xmath92 excluding the code 2 sources from the himf analysis as did @xcite guarantees that more confident detections with well - understood selection criteria are used . it could be argued that the use of sources of code 2 in the analysis could provide value added to the determination of the himf . this is discussed further in [ sec : limits ] . in practice , statistical studies requiring stringent requirements on sensitivity limits should use only code 1 sources and equation [ eq : eqnthresh90c1 ] . with the proper caution associated with the incomplete nature of code 2 sources , the combination of code 1 and code 2 sources and equation [ eq : eqnthresh90c12 ] can be used in studies which can benefit from a larger sample . in both cases , the 50% completeness limit can be considered the `` sensitivity limit '' of the survey , since it is the most relevant completeness limit for the derivation of galaxy statistical distributions , such as the himf and the hi width function . @xcite have shown that adopting a step function cut at the 50% completeness limit of a survey produces approximately the same statistical results as adopting the survey s full completeness function . the 25% completeness limit can be identified with the `` detection limit '' of the survey , that is the integrated flux density level below which a source has only a small chance of being detected and cataloged . we remind the reader that the quoted limits given here refer to the full @xmath77.40 catalog , and hence are representative of the average alfalfa datacube noise properties . however , because of variations in noise among and within grids and because some localized regions are entirely contaminated by rfi , limits on the hi flux density at arbitrary positions ( e.g. upper limits for non - detections ) must be computed individually , by specific inspection of the spectrum noise properties of the data cubes and their associated `` weights grid '' and the continuum maps . it is the availability of such ancillary information which enables the use of the full alfalfa dataset for stacking @xcite to probe statistical ensembles more deeply . as the previous generation blind hi survey , hipass @xcite set the standard for survey completeness ; by design , alfalfa was intended to surpass and supercede hipass . a reasonable comparison of the impact of the different source detection schemes ( including the absolute level of flux density sensitivity ) may be made by comparing the distribution of the highest - mass galaxies in hipass and alfalfa . for example , one might have anticipated that the original hipass peak - flux density detection scheme @xcite could bias the catalog against edge - on ( extremely wide ) profiles at the highest masses , and such a bias could explain the finding of @xcite that the hipass himf underestimated the number density of the highest mass galaxies . figure [ fig : histbigw ] shows a comparison of the distribution of profile widths in @xmath2 ( open histogram ) and hipass ( filled histogram ) , for objects with log @xmath93 . no obvious difference which would explain a lack of high - mass sources in the hipass catalog is apparent . while the peak - flux density threshold detection could introduce such a bias , it is apparent that the matched filtering technique subsequently applied to the hipass dataset recovers high - width objects as does the technique used in alfalfa . instead , we attribute the lack of extremely high - mass sources in the hipass catalog to that survey s limited redshift extent and its lowered sensitivity near its bandpass redshift limit , both of which resulted in inadequate sampled volume and thus an undercounting of the rare , highest mass hi disks . furthermore , because of its lower sensitivity , poorer angular and spectral resolution and source detection scheme , hipass was limited in its ability to probe the very low - mass and narrow - width hi sources . the spectrometer setup employed by hipass yielded a raw resolution of 13.2 km s@xmath27and of 18 km s@xmath27 after hanning smoothing ; the narrowest objects included in the hipass catalog have w@xmath74 = 30 km s@xmath27 . in contrast , alfalfa s velocity resolution is 11 km s@xmath27 after smoothing is applied , and the @xmath2 catalog thus includes sources with extremely narrow velocity widths . although the signal extraction algorithm adopted by @xcite applied a minimum template width of 30 km s@xmath27 , the refined final process of parameter extraction based on individual examination of the emission region permits finer width estimation . in fact , 289 of the extragalactic objects included in table [ tab : catalog ] have @xmath94 km s@xmath27 . figure [ fig : loww ] examines the distributions of low hi mass systems and their profile velocity widths in the two surveys ; alfalfa is clearly superior in its ability to probe the lowest mass systems . this increased sensitivity to very narrow hi line emission enhances alfalfa s ability to probe the lowest hi masses , which in turn robustly constrains the faint - end slope of the hi mass function , @xmath77 . in fact , at the lowest hi masses , @xmath95 8.0 , the hipass catalog includes only 40 objects while the @xmath2 catalog contains 339 . the ability of alfalfa to sample narrower hi line sources is also critical for the derivation of the hi width function and its relation to the halo mass function @xcite . in drawing conclusions from blind hi surveys about the hi - selected population in the local universe , it is critical to understand the biases in the survey due to its sensitivity limits , uncertainties in the hi line flux densities and distances leading to uncertainties in the derived hi masses , and the impact of large - scale structure in the survey volume . @xcite use a subsample of alfalfa hi sources identified in low density environments to establish a standard of normal hi content and performed an analysis of the completeness of the particular version of the alfalfa catalog they used . @xcite ( see also martin 2011 ) provided an overview of important effects that impact the derivation of the himf by two different methods commonly used to derive mass and luminosity functions , namely the 1/v@xmath96 method and the two - dimensional stepwise maximum likelihood ( 2dswml ) method . in the context of applications such as the derivation of the himf by those two methods , we discuss here in greater detail the magnitude and character of @xmath2 survey properties , its limitations and biases . it is particularly important to understand these effects now because we anticipate the ` 100% alfalfa survey ' to be available in the near future . the large increase in the number of galaxies available for that analysis will decrease the statistical uncertainties on the measurements , thus amplifying the relative impact of systematics and biases . additionally , at that stage it will be less practical to create thousands of realizations to help understand the various effects . the results presented in this section will provide a baseline and dictate procedure for the final measurement of the hi mass function from the completed alfalfa survey . selection effects related to the code 2 sources in @xmath2 are poorly determined . because they require redshift information derived from other sources , they are subject to the limitations of the availability of such confirming data . additionally , alfalfa s sensitivity as a function of distance is strongly affected by rfi especially in the frequency range contaminated by the aviation radar at the san juan airport ( 1350 mhz , corresponding to @xmath97 km s@xmath27 ) . for these reasons , @xcite included only objects with code 1 , detected within 15000 km s@xmath27 . yet it may be argued that the additional information contained in code 2 sources , dipping to lower flux limits , could provide additional insight . a first evaluation of the value added to the himf by code 2 sources relates to the observation that most code 2 sources fall near @xmath98 , a region of the himf well sampled by code 1 sources : the value added is likely thus to be negligible . we explore numerically this expectation . because of the requirement that code 2 sources be identified with an oc of known ( prior ) redshift , most often contributed by optical / ir surveys like sdss , those sources may be biased toward overdensities , toward those regions of the local volume that have been included in specific targeted or wide - area redshift surveys , such as the virgo cluster , and in particular toward those regions of the sky that have been covered in the spectroscopic catalogs of the sdss . does the inferred himf change if code 2 sources are included in its derivation ? we account for hi mass and flux density errors by creating 500 realizations of an himf that includes sources of both code 1 and code 2 , and compare those to 100 realizations of the fiducial himf published in @xcite which contained only the code 1 objects . we use the 2dswml method , but do not jackknife resample . as did @xcite , we restrict the analysis to the contiguous areas contained in @xmath2 and limited to c@xmath99 km s@xmath27 . over the same volume , the inclusion of code 2 sources increases the sample size used for this analysis from the 10,021 included in @xcite to 11,177 . figure [ fig : himfc2 ] displays the hi mass function found when code 2 sources are included in the analysis . the parameters of the function are not strongly affected by the inclusion of these sources . we find @xmath100 ( h@xmath101 mpc@xmath102dex@xmath27 ) = 4.8 @xmath54 .3 @xmath103 10@xmath102 , @xmath104(m@xmath105m@xmath106 + 2 @xmath107 h@xmath108 = 9.96 @xmath54 0.02 and @xmath77 = -1.29 @xmath54 0.02 . these correspond to @xmath109 4.1 @xmath54 0.3 @xmath110 h@xmath111 found by integrating the schechter function , or @xmath109 4.2 @xmath54 0.1 @xmath110 h@xmath111 when summing the binned measurements directly . the fiducial himf which includes only code 1 objects as reported by @xcite finds @xmath100 = 4.8 @xmath54 0.3 , @xmath104(m@xmath105m@xmath106 = 9.96 @xmath54 0.02 , @xmath77 = -1.33 @xmath54 0.02 , @xmath112 ( analytical ) = 4.3 @xmath54 0.3 , and @xmath112 ( summed ) = 4.4 @xmath54 0.1 , all in the same units as expressed for the results with both code 1 and 2 sources . encouragingly , these results indicate that the alfalfa survey s detection coding scheme does not systematically exclude significant sources of hi gas energy density in the local universe . rather , the agreement between the code 1 and the code 1 + 2 himfs suggest that our robust understanding of the survey s sensitivity extends to those weaker sources identified as code 2 objects . the only potentially significant impact is on the faint - end of the himf , influencing both the slope and the points there . the slope parameter @xmath77 is flattened in the code 1 + 2 case , though the two values are just barely within 1@xmath113 of each other . in figure [ fig : himfres ] , we compare the residuals ( the best - fit , fiducial himf schechter model , subtracted from the binned data ) for the case where we consider only code 1 objects ( top panel ) and the code 1 + 2 case ( bottom panel ) . the figure clearly demonstrates that the code 1 + 2 himf measured fewer low - mass objects per unit volume , thereby yielding a flatter slope . this is unsurprising for , in comparison to hi surveys , optical surveys undersample dwarf , low surface brightness galaxies . the typical code 2 detection is a galaxy near @xmath114 , and the redshift distribution of code 2 sources lacks the smattering of low redshift objects present in a hi - selected sample . as a result , code 2s add very few additional sources at the lowest redshifts , at which low hi masses are detectable . this is an example of the fact that adding code 2 sources to the sample is more likely to subtract than to add value to the result : `` more is less '' . we note that in the process of source extraction , a second set of marginal hi line detections has been identified which coincide with possible ocs for which no redshift measurement is available . because the probability of these objects is yet too uncertain , they are not included in the current @xmath2 catalog . future followup observations to be made after the main survey is completed will be undertaken to confirm the reality of the hi line detections . this program will contribute additional low hi line flux density sources to the final alfalfa catalog in this region of the sky . ( unfortunately ) we live on a planet occupied by technologically - active humans . figure 6 of @xcite illustrates the relative spectral weight within the 40% alfalfa survey volume as a function of observed heliocentric velocity . a relative weight of 1.0 indicates that the entire surveyed volume was accessible for source extraction and produced high - quality data . as also evident in the deficit of sources near a distance of @xmath0225 mpc seen in figure [ fig : spanplot ] , the faa radar at the san juan airport contaminates the frequencies corresponding to source at @xmath53 between 15000 and 16000 km s@xmath27 , rendering the detection of sources in this range impossible when the transmitter is on . beyond 16,000 km s@xmath27 , alfalfa s sensitivity recovers , but at the corresponding distance , it is sensitive only to the most massive of galaxies . as a result , this distant volume contributes only a small number of galaxies to the overall @xmath2 sample . for these reasons , the analysis of the himf in @xcite neglected galaxies beyond 15000 km s@xmath27 , so that the results would not be influenced by the large spectral weight gap . this exclusion was especially important in the case of the 2dswml method , since the 1/v@xmath96 method allows the inclusion of explicit corrections for known missing volumes . 2dswml , by contrast , determines the shape of the himf by comparing counts in hi mass bins to a built - in description of alfalfa s flux density sensitivity as a function of distance and width . the large gap , which is not anticipated by this approach , may have caused problems in the analysis were those objects included . @xcite felt it was safer to limit the first measurement of the himf to regions where the spectral weights are relatively smooth , that is , to galaxies within 15000 km s@xmath27 . here , we revisit the issue and consider the influence , if any , of including the full redshift extent of the survey in the himf analysis . in particular , we would expect that the increased bin counts at the very highest hi masses may increase the statistical significance of the himf measurement there . such a possibility is of interest because @xcite determined that alfalfa is more sensitive to high - mass galaxies than hipass was , with himf results indicating that previous blind hi surveys have missed a significant percentage of the most massive hi disks . to test this possibility , we calculated the himf using both methods and using all of the code 1 sources out to 18,000 km s@xmath27 . for each method , we created 250 realizations of the survey to account for flux density , distance , and mass errors . the fit parameters and @xmath112 values are displayed in table [ tab : himf ] for both the 1/v@xmath96 and 2dswml methods . it is worth noting that the 2dswml result is distorted , likely because of the influence of the inaccessible volume and the inability of this method to correct for it . in fact , the 2dswml result drastically underestimates @xmath112 , shifts @xmath104 ( m@xmath115/m@xmath106 to a higher value , and flattens out the low - mass slope @xmath77 . on the other hand , the 1/v@xmath96 method continues to function as expected and results in a reliable measurement . both of the 1/v@xmath96 results are consistent with each other , including the measured values for @xmath112 , but the 2dswml method in the presence of the redshift gap performs poorly . this result confirms the decision by @xcite to limit their analysis to the volume within @xmath116 km s@xmath27 . this analysis provides further evidence of the relative strengths and weaknesses of the two available methods for estimation of the himf . while the 2dswml approach provides a powerful statistical tool , it functions as a ` black box ' method that can not be manipulated by additional knowledge of the survey . in some cases , this may be an advantage , but in the case of alfalfa where we have detailed information about the survey volume , the survey sensitivity , and other factors contributing to the himf , the 1/v@xmath96 method provides a clearer path and a more understandable answer . on the low hi mass end , uncertainties in the conversion from hi line flux density to hi mass are the primary source of error on the himf . unlike the practice in the derivation of most himf results in the literature , the error analysis undertaken here and by @xcite has taken this explicitly into account . because the himf is based on binning galaxies by hi mass and then considering each bin as an independent data point , it is not straightforward to carry hi mass uncertainties through analytically . instead , the alfalfa himf s uncertainties due to mass errors are calculated through the creation of many hundreds of realizations , each with randomly assigned mass ( i.e. , distance and flux density ) errors . here , we elaborate further on the distance estimate scheme used in alfalfa , the biases that would be introduced by using alternative schemes ( i.e. , a pure hubble flow model ) and the overall impact of distance and flux density errors on the hi mass estimates used to construct the himf . the distance estimation scheme adopted for alfalfa was described by @xcite and is summarized briefly here . when distances are based on the adopted flow model , we employ the model s error estimates , constrained by the fit of the model to the observed velocity fit . when distances are estimated using pure hubble flow , the error is estimated to be @xmath010% . we fix a minimum error of 163 km s@xmath27 , based on the local velocity dispersion measured by @xcite . to demonstrate the importance of using the full suite of available information to estimate distances , figure [ fig : dist ] compares the primary distances ( used in @xmath2 ) to the values that would be obtained assuming pure hubble flow . in their estimate of galaxy masses for the himf , the hipass team assumed hubble flow . this is not a safe assumption , particularly in the regions of the sky surveyed in @xmath2 . the virgo cluster , in particular , represents a strong deviation from any assumed relationship between distance and recessional velocity . @xcite showed the danger of assuming pure hubble flow , especially because of the small redshifts accessible to blind hi surveys . those authors concluded that the low - mass slope of the hipass himf was underestimated due to neglecting peculiar velocities , and predicted that a survey in the direction of virgo could severely underestimate the low - mass slope . given the large - scale structure in the @xmath2 volume , we would expect the himf to vary strongly if a poor choice of distance estimate were made . in order to test this , we have re - calculated the 2dswml estimate of the himf using pure hubble flow to estimate distances . that is , we converted the observed heliocentric velocities into the cmb rest frame , and then assumed d@xmath117 = @xmath118 , where we adopt the alfalfa standard @xmath119 = 70 km s@xmath27mpc@xmath27 . in this case , we have no ideal estimate of the distance error , and therefore use 10% of the hubble flow distance or the local dispersion value 163 km s@xmath27 , whichever is greater . as usual , flux density errors are also folded into the mass uncertainties . once again , we create 250 realizations to estimate uncertainties . the resulting hi mass function and schechter fit parameters are displayed in figure [ fig : himfhubble ] . as anticipated @xcite , the use of hubble flow has caused a serious underestimate of the faint - end slope @xmath77 . alfalfa s success at robustly measuring the himf depends not only on large sample size over a cosmologically significant volume , but also on the selection of a reasonable model for distance estimation . given the discussion of distance errors and their large impact on the himf and its uncertainties by @xcite , it is reasonable to ask how large are the hi mass errors when both distance and flux density errors in the @xmath2 sample are taken into account . to obtain robust estimates of hi mass errors , we created many thousand realizations of each galaxy in the @xmath2 sample and applied distance and flux density errors . the result , displayed in figure [ fig : massbin ] , compares the himf mass bin galaxies would nominally fall into assuming a perfect measurement of distance and flux density ( along the abscissa ) to the mean mass of the galaxies assigned to that bin once realistic uncertainties are taken into account . the horizontal uncertainties indicate the 1@xmath113 spread of potential ` true ' masses falling into nominally assigned mass bins . from the figure , it is clear that alfalfa s measurement of the himf and @xmath112 is not prone to large uncertainties above @xmath120 . in the mass range of interest to the missing satellites problem , dwarf galaxy studies , and the low hi mass slope of the himf , that is below @xmath120 , galaxies can easily be mis - assigned to bins , even when a realistic distance model is being used . depending on the large - scale structure in the survey volume , this effect would lead to either an over- or under - estimate of @xmath77 . we therefore take great care to account , conservatively , for mass uncertainties . because blind hi surveys are relatively shallow , with alfalfa probing the local universe only out to z @xmath0 0.06 , inhomogeneity in the survey volume has a strong impact on the derived hi mass function . this effect is particularly true in the case of the 1/v@xmath96 method , which is not as robust against large - scale structure , but the 2dswml method is not completely immune from these effects . to test the homogeneity of a sample , the usual statistical test applied is the v / v@xmath96 test @xcite . much like the 1/v@xmath96 method , this test considers the maximum volume out to which each source in a survey can be detected . by comparing the actual volume the source was detected in to the accessible volume , homogeneity in the sample can be evaluated ; the expectation value @xmath121 is 0.5 in a homogeneous volume . in the case of the @xmath2 volume , @xmath121 = 0.45 this indicates that , at 40% completion , the survey does not yet contain enough volume to fully ` smooth out ' the effects of large - scale structure . this is reflected in figure [ fig : vvmax ] where @xmath121 is shown for each bin of hi mass . the most obvious feature in this figure , the dip near log ( m@xmath122/m@xmath106 @xmath123 , is due to overdensities in the sample volume , primarily the virgo cluster . galaxies in those overdensities are found , preferentially , in those regions , rather than filling the full volume where alfalfa s sensitivity could detect them , causing this dip . it is clear that @xmath2 does not , yet , constitute a representative slice of the universe ; as the survey progresses , we anticipate that the full sample will pass the @xmath124 test . another , perhaps more intuitive , way to view the impact of voids and clusters in @xmath2 is to compare the redshift distribution of cataloged galaxies to the prediction based on the survey s selection function ( i.e. , the percentage of galaxies at a given distance that are detectable in alfalfa ) . the selection function is determined by the 2dswml analysis of the himf , and when combined with the measurement of the himf , predicts the redshift distribution for a homogeneously distributed set of galaxies selected from the himf . figure [ fig : zdistr ] compares this expectation to the observations in @xmath2 . the bumps and dips in the histogram represent under- and overdensities , respectively , in the survey volume . for example , the virgo cluster explains the enhancement near 1,000 km s@xmath27 . the pisces - perseus supercluster and its foreground void also make clear imprints on this figure . if @xmath2 does not represent a representative sampling of the universe , then statistical studies of the sample s characteristics , like the himf , may be subject to biases from large - scale structure . because of its size , we can make an assessment by the impact of large scale structure within separate subregions of the catalog . the @xmath2 sample is made up of 3 large , contiguous areas . in the northern galactic hemisphere , @xmath2 covers 07@xmath12530@xmath126 @xmath127 16@xmath12530@xmath126 in two separate blocks , 4@xmath12816@xmath129 and 24@xmath130 28@xmath129 . we refer to these subregions as @xmath131 and @xmath132 , respectively . in the southern galactic hemisphere , @xmath2 covers 22@xmath125 @xmath127 03@xmath125 , 24@xmath13032@xmath129 , referred to as @xmath133 . the entire @xmath2 , combined together , covers enough cosmological volume for the effects of large - scale structure on the derivation of the himf to begin to become minimal , but reducing its coverage further leads to a situation in which the himfs derived for individual sub - regions are strongly affected by over- and under - densities within their volume . figure [ fig : himfsub ] displays the the himfs for the three subregions : @xmath131 , @xmath134 , and @xmath135 , from top to bottom . the fit parameters and values of @xmath112 are given in table [ tab : himfreg ] along with the fiducial 2dswml himf for the entire @xmath2 sample for comparison . the largest by a significant fraction is @xmath131 , and it contributes over 50% of the 10,000 galaxies in @xmath2 . as expected , the himf for this region , when isolated , follows the himf for the sample as a whole . because of the large volume in this region , the himf displayed in the top panel of figure [ fig : himfsub ] is smooth and featureless . in the case of the smaller samples in the middle and bottom panels of figure [ fig : himfsub ] , features due to large - scale structure are clearly visible . because of the inhomogeneity of the surveyed volume , the himfs do not follow the prescribed schechter function . in the case of the @xmath132 subsample , the faint - end slope is better fit on its own , in which case it is measured to be @xmath77 = -1.4 @xmath54 0.1 . in every case , the ` bumps ' and wiggles in the sub - region himfs correspond to the cone diagram distributions in @xcite . in essence , the combination of the alfalfa survey s sensitivity and the scaling of survey volume with redshift leads to preferred distances for each of the hi mass bins in the himf ( or preferred hi masses for every distance in the survey ) . a dip , for example , in the himf corresponds to an overdensity at the preferred distance for those hi mass scales . while the 2dswml method has been designed to be less sensitive to large - scale structure , the volumes of these subregions are too small for these effects to average out . such techniques can only work with the data they are given , but the 1/v@xmath96 approach allows for explicit correction for known structures . when these corrections are included in the 1/@xmath136 analysis of these subregions , the ( unshown ) results are very similar to those provided here . these corrections are based on the iras point source catalog redshift survey ( pscz ; @xcite ) density correction ( see [ lsscorr ] ) , but imperfections in this correction lead to the same bump and dip features . an additional weakness of the 1/v@xmath96 density correction is that the counts can only be increased for galaxies that do end up in the sample , making the correction significantly less useful in voids . by contrast , 2dswml essentially ` self - corrects ' for over- and underdense regions . rather than looking at volumes and scaling counts by 1/v@xmath96 , 2dswml constructs the relationship between bins by scaling the counts themselves and therefore automatically scales the himf downward for regions that are overdense and upward for regions that are underdense . this consideration of subregions within @xmath2 makes clear the impact that large - scale structure can have on blind hi surveys and the importance of cosmologically significant volumes before global conclusions can be drawn . as described in @xcite , the 1/v@xmath96 method can be corrected to account for large - scale structure in the survey volume . essentially , overdense regions are considered to represent more effective volume ( @xmath137 , rather than @xmath138 ) and vice versa for underdense regions , so that galaxies in various environments are weighted appropriately @xcite . while this correction is successful , it does rely on datasets external to the alfalfa survey . in @xcite , the density map derived from the pscz @xcite was used to correct for large - scale structure . however , other options exist , in particular other pscz maps ( smoothed to different levels ) and the density reconstruction derived from the 2mass redshift survey ( 2mrs ; @xcite ) . the large - scale structure correction used has a large influence on the final himf estimate ; a @xmath0 20% effect on the schechter parameters , compared to neglecting the density correction , was reported in @xcite . given the magnitude of the effect , it is important to consider the impact that a different choice would make . in particular , since this portion of himf analysis is likely to always rely on external information , examining it here may be helpful in the future for the 100% alfalfa sample . the parameter of interest reported by pscz is the overdensity @xmath139 , defined relative to the average number density of galaxies found in those surveys : @xmath140 in the case of the alfalfa survey and the himf , we are primarily interested in the average value of @xmath139 interior to each source s maximum detectable distance or , in other words , the average value of @xmath139 in the volume over which the source could have been detected . both the 2mrs and pscz density maps report the value of @xmath139 in equal - volume cells throughout their survey volume . the pscz map was chosen because of its greater sensitivity in the nearby survey regions of @xmath2 , where the himf was especially vulnerable to the impact of large - scale structure . while pscz was a good choice for the analysis of the @xmath2 himf , there are actually several choices of maps available from @xcite , with the primary differences being the smoothing size of each volume cell and the maximum distance out to which the density fields were reconstructed . in @xcite , the chosen map extended to 240 @xmath141 mpc and was smoothed with a gaussian kernel of width 3.2 @xmath141 mpc . the alternative options include a map that extends to only 120 @xmath141 mpc with a 3.2 @xmath141 mpc kernel , and one that extends to 240 @xmath141 mpc with a larger gaussian kernel of 7.7 @xmath141 mpc . the smoothing scale of pscz maps can lead to underestimation of density contrasts . because the primary effect of the large - scale structure correction is on the lowest - mass bins of the himf , it is important to explore and understand the influence of this outside dataset . upon examination of the average interior overdensities determined for @xmath2 in each map , we find that the pscz.240.g3.2 map used in @xcite for the 1/v@xmath96 analysis of the himf represents an extreme estimate of the impact of large - scale structure within the survey volume . it is the most conservative option , given that it attaches lower weight to those galaxies found in nearby overdensities , particularly the virgo cluster , to prevent them from artificially boosting the faint - end slope . in order to quantify the effect of these options on the resulting himf , we use the pscz.120.g3.2 and pscz.240.g7.7 to re - analyze the @xmath2 himf . where the maps do not reach the full redshift extent of the @xmath2 sample , we set the average interior @xmath139 to 0 for galaxies beyond the distance limit . in order to fit schechter function parameters in each case , we use the same uncertainty estimates for each hi mass bin point as presented in @xcite , as the pscz map applied would only change the magnitude of each point and not its fluctuation due to hi mass uncertainties . figure [ fig : himfpscz ] shows the results , focusing on the low - mass end of the himf , since hi mass bins with @xmath142 are not affected by the large - scale structure volume correction . the different large - scale structure corrections function effectively as a scaling in each bin , so that each option follows the fiducial case closely . both pscz.120.g3.2 and pscz.240.g7.7 boost the faint - end slope , indicating that they are overcounting galaxies in the nearby overdensities , namely the virgo cluster . this analysis verifies that pscz.240.g3.2 was the most conservative choice for correcting the 1/v@xmath96 himf for the effects of large - scale structure . the changes to the low - mass slope @xmath77 and the turnover mass m@xmath115 are displayed in table [ tab : himfpscz ] , along with the measured 2dswml parameters for reference . it is clear that the pscz map with the greatest extent and the smallest smoothing radius is most appropriate for estimating the @xmath2 hi mass function . this paper presents the catalogued parameters for 15855 hi line detections extracted from @xmath143 deg@xmath1 of high galactic latitude sky observed by the alfalfa survey . a ( pleasant ) surprise for us has been the higher than expected alfalfa detection rate , 5.6 sources per deg@xmath1 , or , including only the objects that are certainly extragalactic , 5.3 sources per deg@xmath1 . this latter detection value is a factor of 29 times greater than the rate of 0.18 sources per deg@xmath1 achieved by hipass . the characteristic resolution of the alfalfa spectral grids is about 4@xmath144 ; the positions of the hi sources can be determined to an accuracy typically better than 20 . using the publicly available sdss and dss2 imaging datasets , we have assigned probable optical counterparts to more than 98% of the 15041 extragalactic detections and provide a cross - reference to the sdss dr7 photometric and imaging databases . an additional 814 hi line detections can not be identified with stellar counterparts but lie within velocity ranges characteristic of the galactic / circumgalactic hvcs . roughly 3/4 of the optically `` dark '' extragalactic hi sources are located in fields containing galaxies of known optical redshift ; many are likely to be associated with tidal debris fields . we identify four objects as candidate oh megamasers redshifted to @xmath61 ; one of those is a rediscovery of a previously recognized ohm and is associated with a galaxy of the same optical redshift @xcite . future works will explore more systematically the ohm candidates throughout the alfalfa bandpass and also will search for evidence of hi in absorption @xcite . unsurprisingly , a census of the hi bearing population of galaxies in the local universe is strongly biased against galaxies on the red sequence , but some luminous , red galaxies are detected in the hi line . in particular , alfalfa provides a rich sampling of the low - to - moderate density universe at @xmath145 . as a major alfalfa data release , the @xmath2 catalog presented here supercedes the datasets published by our team previously . in particular , the hi line flux densities reported here are based on further improvements in the software used for parameter extraction and increased knowledge of the system performance . the alfalfa reduction pipeline may miss flux for sources which are very large compared to the beam size and offset from the center of the standard grids , but comparison with the hi line flux densities derived from pointed single dish observations and corrected for beam dilution and pointing errors with the ones reported here shows no systematic offsets except for the very largest and very strongest sources . the latter will need to be evaluated on a case - by - case basis in grids produced and analyzed separately from the standard process and where applicable , corrected for sidelobe contamination @xcite . the goals and expectations of the alfalfa survey were outlined in @xcite and survey source sensitivity and reliability was discussed in @xcite . as discussed previously , the integrated hi line flux density threshold of a blind hi survey like alfalfa increases with hi line profile width @xcite . with the availability of the large @xmath2 dataset , we test those expectations and give quantitative descriptions of the completeness and sensitivity of the alfalfa survey as functions of @xmath79 . in addition to the highest quality , highly reliable ( code 1 ) hi detections , the @xmath2 catalog presented in table [ tab : catalog ] includes also sources of lower s / n which coincide in position and redshift with known optical galaxies ( the `` priors '' ) . because the availability of such prior information is highly dependent on the selection functions of other surveys , these additional objects should not be used in studies which require stringent consideration of statistical completeness . however , the vast majority are likely to be valid hi detections and hence they can be included in studies where the number of sources is most critical ( e.g. , peculiar velocity studies ) . future work will be undertaken to confirm these detections and an additional set of low s / n possible detections which coincide with galaxies of unknown redshift . the sensitivity of alfalfa and the thorough understanding of its performance enable a robust measurement of the himf , and in particular , of its faint - end slope @xmath77 and the energy density of neutral hydrogen @xmath112 at @xmath67 = 0 . on the low - mass end of the himf , alfalfa improves on previous blind hi surveys in terms of sample size , angular and spectral resolution , sampling of cosmic volume , and assumptions of pure hubble flow . at the lowest hi masses , alfalfa s finer velocity resolution is an important factor in obtaining a full count of the gas - rich dwarf population . on the high - mass end , previous hi surveys have overlooked the locally - rare population of very massive hi disks . we have evaluated the possible impact on the derived himf of missing sources at both the broad and narrow width ends , particularly in comparison with the hipass catalog @xcite . we conclude that hipass did not recognize the richness of the very high hi mass population , not because it failed to identify the systems with the broadest widths but because it did not have adequate sensitivity at large distances and was limited to only 64 mhz of bandpass . it is alfalfa s combination of sensitivity , spectral and angular resolution , frequency and sky coverage which yields a robust census of the hi bearing population at @xmath146 . with alfalfa still only 40% complete , we have shown that the 2dswml and 1/@xmath136 methods yield results on the himf in good agreement , but that the loss of significant volume in the alfalfa survey beyond 15000 km s@xmath27 reduces the performance of the 2dswml approach if that region is included . a realistic treatment of distance and flux density uncertainties , translated into mass uncertainties , avoids the strong bias in @xmath77 and the shape of the himf introduced by an assumption of hubble flow in the local volume . while @xmath2 does not yet provide a completely representative sampling of the local cosmological volume , our method for including the impact of large - scale structure is a conservative choice , and future data releases from alfalfa will further improve both statistical and systematic uncertainties . we look forward to completing the alfalfa survey . we thank the staff of the arecibo observatory , particularly phil perillat , ganesh rajagopalan , arun venkataraman , hector hernandez , and the telescope operations staff , for their invaluable help in support of the acquisition of the data used to produce this catalog , and tom shannon and adam brazier for their critical support of hardware and database development at cornell . this work has been supported by the nsf grant ast-0607007 to rg and mph and by a brinson foundation grant . tb , dc , glh , sjuh , dk , rk , jm , ao , ro , jr and ew acknowledge support for the undergraduate alfalfa team from nsf grants ast-0724918 , ast-0725267 , ast-0725380 , ast-0902211 and ast0903394 . jlr acknowledges support from nsf ast-000167932 . ks acknowledges support from the natural sciences and engineering research council of canada ( nserc ) . this research has made use of data obtained from or software provided by the us national virtual observatory , which is sponsored by the national science foundation and of montage , funded by the national aeronautics and space administration s earth science technology office , computation technologies project , under cooperative agreement number ncc5 - 626 between nasa and the california institute of technology . montage is maintained by the nasa / ipac infrared science archive . we acknowledge the use of nasa s _ skyview _ facility ( http://skyview.gsfc.nasa.gov ) located at nasa goddard space flight center and the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration . the digitized sky surveys were produced at the space telescope science institute under u.s . government grant nag w-2166 . the images of these surveys are based on photographic data obtained using the oschin schmidt telescope on palomar mountain and the uk schmidt telescope . the plates were processed into the present compressed digital form with the permission of these institutions . the second palomar observatory sky survey ( poss - ii ) was made by the california institute of technology with funds from the national science foundation , the national geographic society , the sloan foundation , the samuel oschin foundation , and the eastman kodak corporation . funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the us department of energy , the nasa , the japanese monbukagakusho , the max planck society and the higher education funding council for england . the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions . the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max planck institute for astronomy , the mpa , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory and the university of washington . rlccrrrrrrrl 331061 & 456 - 013 & 000002.5 + 155220 & 000002.1 + 155254 & 6007 & 260 ( 45 ) & 1.13(0.09 ) & 6.5 & 2.40 & 85.2 & 9.29 & 1 i + 331405 & & 000003.3 + 260059 & 000003.5 + 260050 & 10409 & 315 ( 8) & 2.62(0.09 ) & 16.1 & 2.05 & 143.8 & 10.11 & 1 i + 102896 & & 000006.8 + 281207 & 000006.0 + 281207 & 16254 & 406 ( 17 ) & 2.37(0.12 ) & 11.2 & 2.31 & 227.4 & 10.46 & 1 i * + 102574 & & 000009.1 + 280543 & & -368 & 23 ( 3 ) & 1.29(0.08 ) & 11.2 & 5.05 & & & 9 u * + 102975 & & 000012.3 + 290137 & & -367 & 23 ( 3 ) & 2.85(0.07 ) & 26.7 & 4.69 & & & 9 u * + 102571 & & 000017.2 + 272359 & 000017.3 + 272403 & 4654 & 104 ( 3 ) & 2.00(0.06 ) & 19.0 & 2.29 & 65.9 & 9.31 & 1 i + 102976 & & 000019.0 + 285931 & & -365 & 26 ( 2 ) & 2.53(0.11 ) & 18.3 & 5.76 & & & 9 u * + 102728 & & 000021.2 + 310038 & 000021.4 + 310119 & 566 & 21 ( 6 ) & 0.31(0.03 ) & 7.5 & 1.92 & 9.1 & 6.78 & 1 i + 102575 & & 000028.0 + 280845 & & -371 & 33 ( 7 ) & 0.47(0.03 ) & 8.6 & 2.11 & & & 9 u * + 12896 & 478 - 010 & 000030.1 + 261928 & 000031.4 + 261931 & 7653 & 170 ( 10 ) & 3.14(0.08 ) & 22.0 & 2.44 & 104.5 & 9.91 & 1 i * + 102729 & & 000032.1 + 305152 & 000032.0 + 305209 & 4618 & 53 ( 6 ) & 0.70(0.04 ) & 10.5 & 2.02 & 65.4 & 8.85 & 1 i + 102576 & & 000035.3 + 262712 & & -430 & 21 ( 2 ) & 0.60(0.04 ) & 11.7 & 2.50 & & & 9 u * + 102730 & & 000040.1 + 315610 & 000039.5 + 315618 & 12631 & 79 ( 23 ) & 0.66(0.05 ) & 7.3 & 2.25 & 175.8 & 9.68 & 1 i + 102578 & & 000042.3 + 263311 & & -429 & 22 ( 3 ) & 0.67(0.04 ) & 12.8 & 2.44 & & & 9 u * + 101866 & & 000050.1 + 141612 & 000047.9 + 141639 & 10877 & 291(149 ) & 0.79(0.11 ) & 4.1 & 2.52 & 150.3 & 9.62 & 2 i * + 12901 & 499 - 035 & 000059.5 + 285431 & 000058.9 + 285441 & 6896 & 395 ( 5 ) & 5.03(0.11 ) & 25.2 & 2.24 & 93.7 & 10.02 & 1 i * + 102731 & fgc290a & 000109.3 + 305221 & 000106.4 + 305247 & 7366 & 257 ( 8) & 1.33(0.08 ) & 8.9 & 2.08 & 100.5 & 9.50 & 1 i + 102977 & & 000108.7 + 284738 & & -364 & 22 ( 3 ) & 2.03(0.11 ) & 13.8 & 6.64 & & & 9 u * + 102861 & & 000110.1 + 320425 & & -181 & 22 ( 1 ) & 7.30(0.06 ) & 55.0 & 4.69 & & & 9 u * + 102732 & & 000114.8 + 312218 & 000115.0 + 312227 & 12532 & 292 ( 5 ) & 1.54(0.09 ) & 9.1 & 2.20 & 174.3 & 10.04 & 1 i + 101869 & & 000127.1 + 142431 & 000131.4 + 142427 & 12639 & 183 ( 16 ) & 1.00(0.09 ) & 6.4 & 2.57 & 175.4 & 9.86 & 1 i * + 102733 & & 000129.8 + 311418 & 000130.0 + 311403 & 12581 & 134 ( 12 ) & 1.03(0.08 ) & 8.6 & 2.29 & 175.0 & 9.87 & 1 i + 12911 & n7806 & 000131.5 + 312629 & 000130.1 + 312631 & 4767 & 231 ( 23 ) & 1.40(0.08 ) & 9.4 & 2.19 & 67.5 & 9.18 & 1 i * + 331082 & 433 - 016 & 000134.5 + 150448 & 000134.0 + 150454 & 6368 & 118 ( 8) & 2.72(0.08 ) & 21.4 & 2.60 & 85.9 & 9.67 & 1 i + 748776 & & 000142.4 + 135019 & 000141.3 + 135033 & 6337 & 53 ( 5 ) & 0.65(0.05 ) & 8.7 & 2.27 & 89.9 & 9.09 & 1 i + rccl 102896 & & 1 & in region affected by rfi ; parameters uncertain ; near smaller agc 102897 ( 000005.5 + 281129 , unknown cz ) at 0.7 arcmin + 102574 & & 9 & hvc ; first of two knots near the top of the grid ; see also agc 102575 at 5.1 arcmin + 102975 & & 9 & hvc ; part of filament that stretches through most of this grid + 102976 & & 9 & hvc ; part of a filament that extends beyond this grid into 0004 + 29 + 102575 & & 9 & hvc ; second of two knots near the top of the grid ; see also agc 102574 at 5.1 arcmin + 12896 & & 1 & near agc 331800 ( mcg+04 - 01 - 009 , 0000316 + 261818 , cz=7754 ) at 1.2 arcmin + 102576 & 2- 4 & 9 & compact hvc ; one of two nearby knots ( the other is agc 102578 + 102578 & 2- 5 & 9 & compact hvc ; one of two nearby knots ( the other is agc 102576 ) + 101866 & & 2 & ambiguous oc ; several near including agc 103024 ( 000049.5 + 141532 , unknown cz ) at 1.2 arcmin ; others may be background + 12901 & & 1 & small companion at 0.4 arcmin agc 103021 ( 000057.5 + 285427 , unknown cz ) + 102977 & & 9 & hvc ; faint south end of a filament that stretches through most of this grid + 102861 & & 9 & hvc 110.7 - 29.6 part of nice arc + 12911 & & 1 & multiple system ngc 7805/6 ; ugc 12908 = ngc 7805 group ; blend ? + 101869 & & 1 & agc 101869 ( 000149.5 + 142623 , cz=12568 ) at 5.7 arcmin + 102862 & & 9 & hvc 110.5 - 31.0 part of nice arc + 102978 & & 9 & hvc ; part of filament that stretches through most of this grid + 102735 & & 1 & optical identification with bluer galaxy in pair ; agc 102831 ( 000250.0 + 281725 , unknown cz ) at 0.3 arcmin + 102863 & & 9 & hvc 110.8 - 30.0 part of nice arc + 102979 & & 9 & hvc ; part of filament that stretches through most of this grid + 749126 & & 9 & hvc 1 - 6.04 - 45.19 + 102864 & & 9 & hvc 110.7 - 30.7 part of nice arc + 749127 & & 9 & hvc 105.34 - 47.32 + 102981 & & 1 & oc identified with larger of pair ; second is agc 103015 ( 000250.0 + 281725 , unknown cz ) at 1.6 arcmin + 7 & & 1 & oc identified with larger of pair ; second is agc 100849 ( 000306.3 + 155834 , unknown cz ) at 1.2 arcmin + 100011 & & 2 & poor spatial and spectral definition + rccccrrcc 331061 & 1 & i & 587730775499407375 & 211330582074884096 & 14.77 & 1.59 & 0.02002 & 0.00010 + 331405 & 1 & i & 587740589481525478 & & 15.11 & 1.97 & & + 102896 & 1 & i & 758874370996764887 & & 15.26 & 2.26 & & + 102571 & 1 & i & 758874297994314032 & & 16.10 & 1.39 & & + 102728 & 1 & i & 758874299066483769 & & 18.93 & 2.04 & & + 12896 & 1 & i & 758874370460680283 & & 13.98 & 1.29 & & + 102729 & 1 & i & 758874299066548754 & & 18.32 & 1.43 & & + 102730 & 1 & i & 758874299602960817 & & 16.94 & 1.55 & & + 101866 & 2 & i & 587730773351989400 & 211330580741095424 & 15.15 & 2.48 & 0.03613 & 0.00010 + 12901 & 1 & i & 758874371533308165 & & 13.69 & 2.64 & & + 102731 & 1 & i & 758874372069392715 & & 16.01 & 1.70 & & + 102732 & 1 & i & 758874299603223055 & & 14.91 & 1.93 & & + 101869 & 1 & i & 587727221413707929 & 211330580573323264 & 15.82 & 1.85 & 0.04189 & 0.00010 + 102733 & 1 & i & 758874299603288292 & & 15.85 & 1.82 & & + 12911 & 1 & i & 758874299603222635 & & 13.25 & 3.00 & & + 331082 & 1 & i & 587730774425796793 & 211330582490120192 & 14.87 & 1.46 & 0.02123 & 0.00007 + 748776 & 1 & i & 587730772815184088 & & 16.96 & 1.14 & & + 102734 & 1 & i & 758874372605739306 & & 15.90 & 1.35 & & + 101873 & 1 & i & 587727223561257129 & 211330582536257536 & 16.35 & 2.38 & 0.04254 & 0.00009 + 102735 & 1 & i & 758874299603419848 & & 18.38 & 0.67 & & + 101877 & 1 & i & 587727221413773686 & 211330580648820736 & 16.70 & 1.40 & 0.01734 & 0.00033 + 102980 & 1 & i & 758874371533635834 & & 15.87 & 1.76 & & + 12920 & 1 & i & 758874298531316055 & & 15.13 & 2.04 & & + 100006 & 1 & i & 758874372606001325 & & 14.28 & 2.70 & & + 100008 & 1 & i & 758874372069982254 & & 16.35 & 1.48 & & + rccrrrrrr 102708 & 000337.0 + 253215 & 000336.1 + 253204 & & 0.169 & -1335 & 0.91 & 5.7 & 2.33 + 102850 & 002958.8 + 305739 & 002958.2 + 305832 & & 0.172 & -596 & 0.46 & 6.7 & 2.09 + 181310 & 082311.7 + 275157 & 082312.7 + 275138 & 0.16783 & 0.168 & -1551 & 2.17 & 15.9 & 2.18 + 228040 & 124540.5 + 070337 & 124545.7 + 070347 & & 0.172 & -624 & 0.33 & 5.1 & 2.11 + cccccc 1/v@xmath96 , 15,000 km s@xmath27 & -1.33 ( 0.04 ) & 3.1 ( 0.6 ) & 9.95 ( 0.05 ) & & 4.4 ( 0.1 ) + + 1/v@xmath96 , 18,000 km s@xmath27 & -1.34 ( 0.03 ) & 3.8 ( 0.6 ) & 9.92 ( 0.04 ) & & 4.3 ( 0.1 ) + + 2dswml , 15,000 km s@xmath27&-1.34 ( 0.02 ) & 4.7 ( 0.3 ) & 9.96 ( 0.01 ) & 4.3 ( 0.3 ) & 4.4 ( 0.1 ) + + 2dswml , 18,000 km s@xmath27&-1.26 ( 0.02 ) & 3.4 ( 0.2 ) & 10.00 ( 0.01 ) & 3.0 ( 0.2 ) & 3.1 ( 0.1 ) + cccccc north1&-1.35 ( 0.02 ) & 4.4 ( 0.3 ) & 9.98 ( 0.02 ) & 4.3 ( 0.4 ) & 4.4 ( 0.1 ) + + north2&-1.25 ( 0.04 ) & 5.6 ( 0.6 ) & 9.92 ( 0.02 ) & 4.2 ( 0.5 ) & 4.3 ( 0.2 ) + + south&-1.30 ( 0.04 ) & 4.1 ( 0.5 ) & 9.96 ( 0.3 ) & 3.6 ( 0.5 ) & 3.5 ( 0.2 ) + + whole @xmath2&-1.34 ( 0.02 ) & 4.7 ( 0.3 ) & 9.96 ( 0.01 ) & 4.3 ( 0.3 ) & 4.4 ( 0.1 ) +

we present a current catalog of 21 cm hi line sources extracted from the arecibo legacy fast arecibo l - band feed array ( alfalfa ) survey over @xmath02800 deg@xmath1 of sky : the @xmath2 catalog . covering 40% of the final survey area , the @xmath2 catalog contains 15855 sources in the regions @xmath3 r.a . @xmath4 , @xmath5 dec . @xmath6 and @xmath7 dec . @xmath8 and @xmath9 r.a . @xmath10 , @xmath11 dec . @xmath6 and @xmath7 dec . @xmath12 . of those , 15041 are certainly extragalactic , yielding a source density of 5.3 galaxies per deg@xmath1 , a factor of 29 improvement over the catalog extracted from the hi parkes all sky survey . in addition to the source centroid positions , hi line flux densities , recessional velocities and line widths , the catalog includes the coordinates of the most probable optical counterpart of each hi line detection , and a separate compilation provides a crossmatch to identifications given in the photometric and spectroscopic catalogs associated with the sloan digital sky survey data release 7 . fewer than 2% of the extragalactic hi line sources can not be identified with a feasible optical counterpart ; some of those may be rare oh megamasers at 0.16 @xmath130.25 . a detailed analysis is presented of the completeness , width dependent sensitivity function and bias inherent of the @xmath2 catalog . the impact of survey selection , distance errors , current volume coverage and local large scale structure on the derivation of the hi mass function is assessed . while @xmath2 does not yet provide a completely representative sampling of cosmological volume , derivations of the hi mass function using future data releases from alfalfa will further improve both statistical and systematic uncertainties .

the standard braunstein - kimble cv teleportation protocol , usually described in the domain of quantum optics by the wigner function formalism @xcite , can be expressed equivalently in different mathematical approaches @xcite ( for a comprehensive review on cv quantum teleportation and information processing , see ref . @xcite ) . following a recent proposal , a very convenient description of the protocol can be given in terms of the characteristic functions of the various quantum states involved , including input , resource , and teleported states @xcite . denoting by @xmath0 and @xmath1 , respectively , the single - mode input state to be teleported and the associated characteristic function , and by @xmath2 and @xmath3 , respectively , the entangled two - mode resource and its characteristic function , it can be shown that the characteristic function @xmath4 of the teleported state takes the factorized form @xcite : @xmath5 a commonly used measure of the success probability of a teleportation protocol is the fidelity of teleportation , i.e. @xmath6 $ ] . in the characteristic - function formalism , the fidelity reads @xmath7 in the following we will adopt eq . ( [ fidelitychi ] ) to analyze the efficiency of the cv teleportation protocol . + in ref . @xcite , the squeezed bell - like states have been exploited as non - gaussian entangled resources optimizing the teleportation fidelity for various sets of input states . the general expression of squeezed bell - like states reads : @xmath8^{-1/2 } s_{12}(\zeta ) \{c_{1 } |0,0 \rangle_{12 } + e^{i \theta } c_{2 } \label{squeezbell}\ ] ] where @xmath9 is the two - mode squeezing operator , @xmath10 , @xmath11 is a two - mode fock state , and the @xmath12s are real constants . a convenient parametrization is @xmath13 and @xmath14 . at fixed squeezing parameter @xmath15 , and at fixed phases @xmath16 and @xmath17 , the entangled resource eq . ( [ squeezbell ] ) is endowed with an additional phase @xmath18 , that can be freely adjusted and thus used to maximize the fidelity @xmath19 : @xmath20 such an optimization yields a remarkable enhancement in the success probability of teleportation for various input states @xcite . the exact analytical expression @xmath21 for the fidelity of teleportation of input coherent states using optimized squeezed bell - like resources , i.e. eq . ( [ fidoptsb ] ) , is reported in appendix [ appendix ] . the optimal phase @xmath22 , maximizing the fidelity @xmath21 is given by : @xmath23 + it is now important to observe that the bell - like contribution in eq . ( [ squeezbell ] ) can be obtained from the truncation at first order @xmath24 of twin - beam states , i.e. @xmath25 , with @xmath26 . this fact can be more clearly understood if one chooses , in eq . ( [ squeezbell ] ) , the parametrization @xmath27 and @xmath28 ( with @xmath29 ) . thus , although being a non - gaussian state , the squeezed bell - like state can be extracted by a gaussian squeezed state by means of a truncation procedure ; in other words , it is a squeezed truncated - twb . however , the connection of entangled non - gaussian resources with the gaussian twin beams may break down if we further generalize the squeezed bell - like state ( [ squeezbell ] ) to the following class of non - gaussian states , the two - mode squeezed symmetric superposition of fock states : @xmath30^{-1/2 } s_{12}(\zeta)\{c_{1 } |0,0 \rangle_{12 } + e^{i \theta_{2 } } c_{2 } \label{squeezssf}\ ] ] where @xmath12 and @xmath31 are real constants and phases , respectively . with respect to state ( [ squeezbell ] ) , state ( [ squeezssf ] ) is not constrained to belong to the class of squeezed truncated - twb . namely , thanks to the additional four - photon term in the superposition , the class of squeezed symmetric superposition of fock states includes the squeezed truncated - twb as a special realization . now , it is very interesting to determine the subset of states @xmath32 corresponding to an optimal fidelity of teleportation . obviously , it is expected that the exploitation of the resource ( [ squeezssf ] ) will also allow a further optimization of the teleportation protocol . we compute the teleportation fidelity for an input coherent state @xmath33 and an input single - photon fock state @xmath34 . a convenient parametrization of the @xmath12s in eq . ( [ squeezssf ] ) is provided by the hyperspherical coordinates in three dimensions : @xmath35 , @xmath36 , @xmath37 . the two - mode characteristic function @xmath38 of the entangled resource ( [ squeezssf ] ) can be calculated as @xmath39 = \,_{ssf}\langle\psi|d_{1}(\alpha_{1})d_{2}(\alpha_{2})|\psi\rangle_{ssf } \ , , \label{chissf}\ ] ] where @xmath40 is the glauber operator for the mode @xmath41 . the explicit expression for eq . ( [ chissf ] ) can be easily written by making use of the two - mode bogoliubov transformations @xmath42 and of the relation @xmath43 where @xmath44 denotes the associated laguerre polynomial . the characteristic functions @xmath45 and @xmath46 corresponding to an input coherent state @xmath33 and to an input single - photon fock state @xmath34 , respectively , are given by : @xmath47 } \ ; , \label{chicohin } \\ \chi_{f}(\alpha_{in } ) & = & e^{-\frac{1}{2}|\alpha_{in}|^{2 } } \ , ( 1-|\alpha_{in}|^{2 } ) \ ; . \label{chifockin}\end{aligned}\ ] ] given eqs . ( [ chissf ] ) , ( [ chicohin ] ) , and ( [ chifockin ] ) , and using eq . ( [ telepchiout ] ) , the fidelities , defined by eq . ( [ fidelitychi ] ) , for the two different input states can be analytically computed . the resulting expressions are functions of the parameters @xmath15 , @xmath48 , @xmath49 , @xmath50 , @xmath51 , @xmath52 , but we do not report them , as they are too long and cumbersome . at fixed squeezing @xmath15 , with maximally fixed phase @xmath16 , the optimized fidelity is defined as @xmath53 where @xmath54 denotes the set of independent free parameters @xmath55 . thus , once the class of single - mode input states is specified , the optimization procedure of the available free parameters , in order to maximize the fidelity , leads to a _ sculpturing of the teleported output state_. we compute the optimal fidelities ( [ optimalfid ] ) , for the above two classes of input states , by numerical maximization of the analytical expressions of @xmath56 . first , we notice that the phases @xmath50 and @xmath52 can be set to zero as they do not contribute to an improvement of @xmath56 . the procedure leads to the following remarkable result : optimization over the set @xmath54 yields maximal parameters compatible with _ a squeezed truncated - twb as optimal entangled non - gaussian resource_. thus , the optimal non - gaussian resources for cv teleportation form a subset of the class ( [ squeezssf ] ) of the form : @xmath57^{-1/2 } s_{12}(-r)\{|0,0 \rangle_{12 } + \tanh s \label{squeezssf'}\ ] ] where , correspondingly to eq . ( [ optimalfid ] ) , the real parameter @xmath58 is fixed to the maximal value @xmath59 , through the numerical maximization of the fidelity , i.e. @xmath60 . in fig . [ figfidsculptelep ] , we show the behavior of @xmath61 as a function of @xmath15 for the input states @xmath33 ( panel i ) and @xmath34 ( panel ii ) , with a squeezed symmetric superposition of fock states as entangled resource . for comparison , also the curves @xmath61 , corresponding to the teleportation with a squeezed state and with a squeezed bell - like state , are plotted . , as a function of the squeezing parameter @xmath15 , with @xmath16 . in panel i , we report the fidelity for the teleportation of input coherent states @xmath33 using different two - mode entangled resources : @xmath62 gaussian squeezed state resource ( full line ) ; @xmath63 squeezed bell - like state ( dashed line ) ; @xmath64 squeezed symmetric superposition of fock states ( long - dashed line ) . in panel ii , we report the fidelity for the teleportation of input single - photon fock states @xmath34 using different two - mode entangled resources : @xmath62 gaussian squeezed state resource ( full line ) ; @xmath63 squeezed bell - like state ( dot - dashed line ) ; @xmath64 squeezed symmetric superposition of fock states ( double - dot - dashed line ) . in plot i the value of @xmath65 is arbitrary . the insets in both panels give the maximal values of the parameter @xmath59 as a function of @xmath15 for given entangled resource and fixed input state . the plotstyles are chosen as specified above.,width=642 ] we observe that , as expected , a further enhancement of the fidelity is obtained , compared to the case discussed in ref . + in ref . @xcite , in order to understand the properties of the optimized resources , three quantities have been investigated : the von neumann entropy @xmath66 , in order to quantify the amount of entanglement present in the resource ; the non - gaussianity @xmath67 , in order to provide a measure of the non - gaussian character of the resource ; and the squeezed - vacuum affinity @xmath68 , in order to determine the degree of resemblance of the resource to a two - mode squeezed vacuum . in fig . [ figvonneumstrunctwb ] , we plot @xmath66 for the optimized non - gaussian entangled resources eqs . ( [ squeezbell ] ) and ( [ squeezssf ] ) , and for the twin beam . it can be observed that all the curves exhibit very similar behaviors . at fixed @xmath15 , and for a given input state , the optimized squeezed symmetric superposition of fock states is the most entangled state , and the teleportation of a single - photon requires more entanglement than the teleportation of a coherent state . however , we notice that the behavior of the von neumann entropies is in agreement with the behavior of the optimal fidelities ; in fact , at fixed input state , a non - gaussian resource with a higher teleportation fidelity is associated with a higher amount of entanglement content for any @xmath15 .. as a function of @xmath15 , with @xmath59 ( see the insets in fig . [ figfidsculptelep ] ) , corresponding to the following entangled resources : squeezed symmetric superposition of fock states , squeezed bell - like states , and two - mode squeezed states , optimized for the teleportation of input coherent states @xmath33 , and input fock states @xmath34 . the plot styles are chosen as specified in fig . [ figfidsculptelep].,width=302 ] next , we examine the behavior of the measure of non - gaussianity @xmath67 , which is defined as the distance between the considered state @xmath69 and a reference gaussian state @xmath70 , according to the following relation : @xmath71+tr[\rho_{g}^{2}]-2tr[\rho \rho_{g}]}{2 tr[\rho^{2 } ] } , \label{nongaussianity}\ ] ] where the gaussian state @xmath70 is completely determined by fixing for it the same covariance matrix and the same first order statistical moments of the quadrature operators associated to state @xmath69 @xcite . in fig . [ fignongaussaffstrunctwb ] panel i , we plot @xmath67 for the entangled resources eqs . ( [ squeezbell ] ) and ( [ squeezssf ] ) , optimized for teleportation . looking at the behaviors of @xmath67 corresponding to teleportation of the same input states , we see that , both in the instance of input coherent states and in the instance of input fock states , the curves cross at a certain value of @xmath15 . therefore , according to the particular adopted measure of non - gaussianity ( [ nongaussianity ] ) , a non - gaussian resource with a higher teleportation fidelity does not exhibit , in general , a higher non - gaussian character for any @xmath15 . ( panel i ) and of the squeezed - vacuum affinity @xmath68 ( panel ii ) as a function of @xmath15 , with @xmath59 ( see the insets in fig . [ figfidsculptelep ] ) , corresponding to the following entangled resources : squeezed symmetric superposition of fock states , and squeezed bell - like states , optimized for the teleportation of input coherent states @xmath33 , and input fock states @xmath34 . the plotstyles are chosen as specified in fig . [ figfidsculptelep ] . , width=642 ] in order to quantify how a given state resembles a squeezed vacuum , a measure of the _ squeezed - vacuum affinity _ can be introduced @xcite . such a quantity is defined as the following maximized overlap : @xmath72 where @xmath73 is a two - mode squeezed vacuum with real squeezing parameter @xmath74 , and @xmath75 is any entangled two - mode resource that depends uniquely on the squeezing @xmath15 as the only free parameter . in fig . [ fignongaussaffstrunctwb ] panel ii , we plot @xmath68 for the entangled resources eqs . ( [ squeezbell ] ) and ( [ squeezssf ] ) , optimized for teleportation . at fixed input states , the optimized squeezed symmetric superposition of fock states exhibits a greater squeezed - vacuum affinity for any @xmath15 with respect to the squeezed bell - like state . this result is in complete agreement with the behaviors of the corresponding optimal teleportation fidelities and von neumann entropies . therefore , comparing the behaviors associated with resources optimized for the efficient teleportation of the same input states , we observe that the following hierarchy can be established : the enhancement of the teleportation fidelity corresponds to an enhancement of the entanglement content , and to an enhancement of squeezed - vacuum affinity . this hierarchy is not respected by the measure of non - gaussianity @xmath67 . this result shows that such a measure must be used and interpreted with care , case by case . as expected , in the plots of fig . [ fignongaussaffstrunctwb ] panel i and panel ii , for sufficiently large @xmath15 , the curves , associated with the same resource , go to the same asymptotic saturation value . + the following conclusions can be drawn from the above discussion : the squeezed truncated twin beam , with truncation at @xmath76 ( four - photon term ) , is an optimal entangled non - gaussian resource suitable for sculptured quantum teleportation . we argue that this result can be generalized to any order @xmath77 in expansions of the squeezing operator at finite order . we thus conjecture that , in general , efficient entangled non - gaussian resources for cv quantum teleportation should be characterized by a suitable balance between the entanglement content and the degree of affinity to squeezed vacuum states . by replacing in eq . ( [ squeezbell ] ) the two - photon term @xmath78 with the two - mode symmetric coherent state @xmath79 , with @xmath80 complex amplitude @xmath81 , the squeezed bell - like state ( [ squeezbell ] ) is converted in the following squeezed cat - like superposition : @xmath82 where the normalization factor is @xmath83 . we remark that the possibility of engineering states of the form @xmath84 is confirmed by the recent successful experimental realization of their single - mode equivalent @xcite . the characteristic function associated with eq . ( [ squeezcat ] ) , @xmath85 can be easily calculated , yielding a simple algebraic sum of gaussian functions . we can thus compute analytically the fidelity of teleportation using squeezed cat - like states ( [ squeezcat ] ) as entangled resource , limiting the analysis , for simplicity , to the teleportation of input coherent states @xmath33 . the maximization procedure is discussed in appendix [ appendix ] , including the analytical expression for the fidelity . at fixed squeezing @xmath15 , the optimal fidelity can be obtained by maximizing on the real amplitude @xmath86 , as the only free parameter @xmath87 , as a function of the squeezing parameter @xmath15 , with @xmath88 . the fidelity corresponds to the teleportation of input coherent states @xmath33 with a squeezed cat - like state as entangled resource ( dotted line ) . for comparison , we plot the fidelities obtained with a squeezed vacuum resource(full line ) and with a squeezed bell - like resource ( dashed line ) . the inset gives the maximal value of the parameter @xmath89 as a function of @xmath15.,width=302 ] in fig . [ figfidsculptelep2 ] , we show the behavior of @xmath61 as a function of @xmath15 for the input states @xmath33 , with an optimized squeezed cat - like state as entangled resource . for comparison , also the optimal fidelities , corresponding to the teleportation with gaussian squeezed state resources and with squeezed bell - like states , are plotted . optimized squeezed cat - like non - gaussian resources yield a significant improvement of the fidelity with respect to the two - mode squeezed vacuum , but are less efficient than squeezed bell - like resources . this result is in agreement with the behaviors of the von neumann entropy , the non - gaussianity , and the squeezed - vacuum affinity , reported in figs . [ figvonneumsqcat ] , and [ figdngaffsqcat ] panel i and panel ii , respectively . in fact , the optimized squeezed cat - like superposition is less entangled , less non - gaussian ( according to the particular measure @xmath67 ) , and less affine to the squeezed - vacuum , than the optimized squeezed bell - like state . in particular , the optimized squeezed cat - like resource exhibits a marked reduction of @xmath68 , if compared with the corresponding plots shown in fig . [ figdngaffsqcat ] panel ii ; in fact , @xmath68 keeps lower than @xmath90 , for any @xmath15 . as a function of @xmath15 , with @xmath89 ( see the inset in fig . [ figfidsculptelep2 ] ) for the optimized squeezed cat - like state ( dotted line ) . for comparison , we also plot the quantities @xmath66 corresponding to a squeezed vacuum ( full line ) , and to a squeezed bell - like state ( dashed line).,width=302 ] ( panel i ) and of the squeezed - vacuum affinity @xmath68 ( panel ii ) as a function of @xmath15 , with @xmath89 ( see the inset in fig . [ figfidsculptelep2 ] ) for the optimized squeezed cat - like state ( dotted line ) . for comparison , in panel i we plot @xmath67 for the optimized squeezed bell - like state ( dashed line).,width=642 ] in this section , we extend the procedures described above to the instance in which , while keeping the protocol ideal , the entangled resources are affected by thermal noise . in order to describe how a mixed resource arises from a pure one when a thermal background is turned on , we consider fields corresponding to independent contributions and converging on a common spatial volume . we follow a very effective scheme proposed by glauber , based on the @xmath91-function formalism @xcite . the superposition of two independently generated fields , both admitting a @xmath91 representation , is simply the convolution of the individual @xmath91-functions : @xmath92 as demonstrated in ref . @xcite ( see also ref . @xcite ) , if _ at least one _ of the independent fields admits a valid @xmath91 representation , then the normally ordered quantum characteristic function @xmath93 of the superposition is given by the multiplication of the individual characteristic functions : @xmath94 thus , the relation ( [ charfuncmultipl ] ) can be applied in the case of the superposition of a generic field @xmath69 with a single - mode thermal field @xmath95 whose normally ordered characteristic function is a gaussian function of the form @xmath96 where @xmath97 denotes the average number of thermal photons . generalizing the formula ( [ charfuncmultipl ] ) to the two - mode instance , we consider the superposition of a two - mode non - gaussian state @xmath98 and a two - mode thermal state @xmath99 . next , we specialize the analysis to the two - mode squeezed bell - like state and to the two - mode squeezed cat - like state , given by eqs . ( [ squeezbell ] ) and ( [ squeezcat ] ) , respectively . the resulting field is then described by the symmetrically ordered characteristic function @xmath100 where @xmath101 is the thermal parameter associated with mode @xmath41 @xmath102 , and @xmath103 denotes the characteristic function associated with the considered non - gaussian state . the state described by eq . ( [ chingth ] ) is mixed as soon as @xmath104 . of course , for @xmath105 we obtain the characteristic function for the superposition of a two - mode squeezed vacuum state @xmath106 and a two - mode thermal state @xmath107 . as in the previous sections , we can now exploit the mixed non - gaussian resource , described by the characteristic function ( [ chingth ] ) , and use eq . ( [ chicohin ] ) to calculate the fidelity of teleportation , for instance in the case of input coherent states . the explicit analytical expressions for the fidelities of teleportation of input coherent states either with mixed squeezed bell - like resources or with mixed squeezed cat - like resources , i.e. @xmath108 and @xmath109 , are reported in appendix [ appendix ] . in the case of mixed squeezed bell - like states , at given thermal numbers @xmath110 , @xmath111 , and fixed squeezing parameter @xmath15 , the optimal fidelity is defined as : @xmath112 and the optimal angle @xmath22 is easily computed : @xmath113 in the case of mixed squeezed cat - like states , at given thermal numbers @xmath110 , @xmath111 , and fixed squeezing parameter @xmath15 , the optimal fidelity is defined as : @xmath114 in eq . ( [ fidthscoptim ] ) the maximization can be performed numerically . in fig . [ figfidtelepmixedng ] , we plot the optimal fidelities for the mixed non - gaussian resources as a function of squeezing for various choices of the thermal parameters @xmath115 . , associated with the teleportation of input coherent states @xmath33 with mixed squeezed bell - like resources ( panel i ) and mixed squeezed cat - like resources ( panel ii ) , as a function of the squeezing @xmath15 , for several choice of the thermal parameters @xmath115 . the curves representing the fidelities associated with mixed non - gaussian entangled resources are plotted with the following plot style : @xmath116 ( full black line ) , @xmath117 ( long - dashed line ) , @xmath118 ( double - dot dashed line ) , and @xmath119 ( dotted line ) . for comparison , in both panels , we also plot the fidelities associated to the mixed squeezed vacuum entangled resources , with @xmath116 ( full gray line ) , @xmath117 ( dashed line ) , @xmath118 ( dot - dashed line ) , and @xmath119 ( long - dotted line ) . the inset in panel ii gives the maximal value of the parameter @xmath89 as a function of @xmath15 , with @xmath120 ( same plot style as for the fidelities).,width=642 ] we observe that , as expected , the fidelity decreases for increasing @xmath97 : the thermal noise sensibly reduces the success probability of teleportation . however , at fixed squeezing and thermal parameters , the non - gaussian resources always perform better than the gaussian twin beam . moreover , also in the mixed instance , mixed squeezed bell - like states perform better than mixed squeezed cat - like states . importantly , for the chosen realistic values of @xmath97 , the fidelity associated with the mixed non - gaussian resources never drops below the threshold of classical teleportation with maximal fidelity @xmath121 . on the other hand , one can look for the threshold value @xmath122 such that , at fixed @xmath15 , one has @xmath123 . [ fignthclassicalfid ] we plot @xmath124 as a function of @xmath15 . as a function of @xmath15 , for mixed squeezed bell - like resources ( dashed line ) , mixed squeezed cat - like resources ( dotted line ) , and mixed squeezed vacuum ( full line).,width=302 ] we see that the threshold value @xmath124 associated to non - gaussian resources is larger than that associated to gaussian ones , and in any case sensibly larger than the reasonable realistic values we have considered . on the contrary , the fidelity associated with the mixed squeezed vacuum state , at low values of @xmath15 , falls below the classical threshold , unless optimization is performed on the local degrees of squeezing , without modifying the braunstein - kimble protocol @xcite . we have analyzed the performance of the braunstein - kimble continuous - variable quantum teleportation protocol using two particular classes of non - gaussian resources , i.e. the squeezed truncated twin beam and the squeezed cat - like states . the extra freely tunable parameters admitted by these resources in comparison to standard gaussian resources allow for an optimization procedure in order to maximize the fidelity of teleportation . for a given set of input states , such a procedure yields a sculptured output teleported state . in order to interpret the results and explain the enhancement of the fidelity , we have characterized and compared resources according to their degrees of entanglement , non - gaussianity , and squeezed - vacuum affinity . we have shown that the optimized non - gaussian resources possess both a high entanglement content and a high degree of squeezed - vacuum affinity . finally , we have studied the efficiency of cv teleportation of input coherent states using , as entangled resources , mixed non - gaussian states of the radiation field . we have shown that , although the thermal noise strongly affects the success probability of teleportation , the resources provided by the optimized mixed - state extension of pure non - gaussian resources guarantee sufficiently high fidelities , for realistic values of the average thermal photon number @xmath97 and of the squeezing @xmath15 . in particular , they assure better performances with respect to the mixed extensions of pure gaussian twin - beam states . + the present analysis can be generalized by considering more general non - gaussian resources and fully realistic situations that involve simultaneously nonideal bell measurements and decoherence induced by propagation in noisy channels . a further goal will be the study of optimal teleportation with non - gaussian resources of two - mode and multimode states @xcite , and the optimization of teleportation protocols with non - gaussian resources with respect to _ local _ parameters and properties , extending the existing schemes for the local optimization of gaussian resources , that yield a one - to - one correspondence between the content of entanglement present in a resource and maximal nonclassical output fidelity @xcite . here we report the analytical expressions of the fidelities of teleportation of input coherent states with the pure and the mixed non - gaussian squeezed bell - like states as resources ( we do not report the whole calculation that is straightforward but somewhat cumbersome ) . the fidelity @xmath21 for the squeezed bell - like state ( [ squeezbell ] ) reads : @xmath125 for @xmath105 , the state @xmath126 reduces to the twin beam , and eq . ( [ fidoptsb ] ) gives the well known formula : @xmath127 the fidelity @xmath108 for the mixed squeezed bell - like state reads : @xmath128\sin 2\delta}{e^{-2r}[1+e^{2r}f_{th}]^{3 } } \ , , \label{fidsbmixed}\ ] ] where we have defined @xmath129 of course , it is easy to check that , in the absence of thermal fields , i.e. @xmath130 , we get @xmath131 , and the fidelity @xmath132 reduces to @xmath19 . notice that , in the limit of large thermal parameter , the fidelity of teleportation ( [ fidsbmixed ] ) is strongly suppressed and vanishes asymptotically . next we give the analytical expressions of the fidelities of teleportation of input coherent states with pure and mixed non - gaussian squeezed cat - like states as resources . by putting , as usual , the phases @xmath133 and @xmath48 to zero , the fidelity for the squeezed cat - like state ( [ squeezcat ] ) reads : @xmath134 it is worth noticing that , for @xmath135 and/or @xmath136 , the fidelity ( [ fidelitysqueezkat ] ) reduces to the well known expression @xmath137 , holding for a twin beam resource . a preliminary numerical optimization procedure for eq . ( [ fidelitysqueezkat ] ) allows to fix the parameters @xmath138 and @xmath18 to the values @xmath139 and @xmath140 , leading to a simplification of the fidelity to the formula : @xmath141 thus , at fixed squeezing @xmath15 , the maximization can be carried out on the real amplitude @xmath86 , as the only free parameter , i.e. @xmath142 . in the particular instance of zero squeezing @xmath143 , it is easy to find value of the optimal fidelity @xmath144^{-1}\simeq 0.6035 $ ] , for @xmath145 . considering the mixed instance , at fixed angle @xmath140 and phases @xmath16 , @xmath17 , @xmath146 , the fidelity @xmath109 for the mixed squeezed cat - like state reads : @xmath147 where @xmath148 is defined in eq . ( [ fth ] ) .

we investigate continuous variable ( cv ) quantum teleportation using relevant classes of non - gaussian states of the radiation field as entangled resources . first , we introduce the class two - mode squeezed symmetric superposition of fock states , including finite truncations of twin - beam gaussian states as special realizations . these states depend on a set of free independent parameters that can be adjusted for the optimization of teleportation protocols , with an enhancement of the success probability of teleportation both for coherent and fock input states . we show that the optimization procedure reduces the entangled resources to truncated twin beam states , which thus represents an optimal class of non - gaussian resources for quantum teleportation . we then introduce a further class of two - mode non - gaussian entangled resources , in the form of squeezed cat - like states . we analyze the performance and the properties of such states when optimized for ( cv ) teleportation , and compare them to the optimized squeezed bell - like states introduced in a previous work @xcite . we discuss how optimal resources for teleportation are characterized by a suitable balance of entanglement content and squeezed vacuum affinity . we finally investigate the effects of thermal noise on the efficiency of quantum teleportation . to this aim , a convenient framework is to describe noisy entangled resources as linear superpositions of non - gaussian state and thermal states . although the presence of the thermal component strongly reduces the teleportation fidelity , noisy non - gaussian states remain preferred resources when compared to noisy twin - beam gaussian states . in recent years , much effort has been dedicated to the theoretical investigation and experimental production of highly nonclassical , non - gaussian states of the radiation field @xcite . such states are endowed with enhanced nonclassical properties , such as entanglement and negative quasi - probability phase - space distribution , that can be useful for improving the efficiency of quantum information and communication protocols @xcite . importantly , a general , rigorous result states that , at fixed covariance matrix , some of these desirable properties are minimized by gaussian states , which can then be considered as extremal `` benchmark '' or lowest - threshold states for quantum information applications @xcite . therefore , at least in principle , the exploitation of non - gaussian states and non - gaussianity allows to quantitatively improve the performance of several quantum information protocols . for instance , non - gaussian cloners of coherent states turn out to be optimal with respect to gaussian ones , exhibiting an enhanced single - clone fidelity @xcite . concerning quantum teleportation with continuous variables ( cv ) , the success probability of cv teleportation can be greatly increased by using entangled non - gaussian resources , degaussified through conditional measurements , consisting of photon - subtraction @xcite . in particular , as shown in ref . @xcite , the implementation of simultaneous phase matched multiphoton processes and conditional measurements can be used to introduce a general class of two - mode non - gaussian entangled states , in the form of squeezed bell - like states depending on an extra free , experimentally adjustable parameter . the optimization on this additional free parameter allows a remarkable increase of the teleportation fidelity for several relevant input states , compared both to gaussian and other degaussified resources @xcite . from the point of view of quantum state engineering , several schemes for the generation of non - gaussian states have been proposed @xcite , and some of them , based on photon addition or subtraction , have been experimentally implemented to obtain photon - added and photon - subtracted single - mode states from gaussian inputs @xcite . more recently , very advanced experimental realizations have been reported @xcite . first , enhancement of entanglement and production of states with negative two - mode wigner function have been successfully obtained via photon subtraction of a single delocalized photon from an initial two - mode gaussian state @xcite . remarkably , the photon - addition / subtraction operations , performed on a thermal light field , have also led to the demonstration of the commutation relation rules , one of the constitutive relations of quantum mechanics @xcite . moreover , using homodyne detection and fock - state resources , the experimental generation of large squeezed schrdinger cat states has been achieved @xcite . these optical cat - like states are of particular importance because they are strongly resilient against decoherence @xcite . + in the present work , we aim to extend the results obtained in ref . @xcite by generalizing the class of squeezed bell - like states and by proposing an alternative class of non - gaussian superposition states , and by showing that these two classes of entangled non - gaussian resources allow for the optimization of cv quantum teleportation . specifically , we investigate cv teleportation with two different classes of two - mode entangled resources : squeezed symmetric superpositions of fock states , and squeezed cat - like states . these two classes share two common features : both are associated to the squeezing operator as cv generator of quantum correlations , and for a trivial choice of the parameters , both reduce to the gaussian twin beam ( twb ) . the last part of the work will be dedicated to the investigation of the effects of the presence of thermal noise on the performance of two - mode non - gaussian states used as resources for quantum teleportation . in particular , we will consider the non - gaussian resources obtained by superimposing squeezed bell - like states or squeezed cat - like states and two - mode thermal states . non - gaussian states and thermal states are assumed to be independently generated , and overlap in a common spatial volume @xcite . due to the thermal contribution , the ensuing mixed states and their correlation properties are modified by the presence of thermal photons . we limit the discussion to the situation of ideal teleportation protocol , i.e. ideal bell measurements and decoherence - free propagation in communication channels . detailed analysis in the instance of the most general realistic situation , including various sources of noise , will be discussed elsewhere . + the paper is organized as follows . in section [ secsqueeztrunctwb ] , we briefly recall the mathematical machinery needed for the description of the ideal braunstein - kimble cv teleportation protocol in the characteristic function formalism . next , we introduce the class of squeezed symmetric superposition of fock states , and apply to them the optimization procedure defined in ref . @xcite . in section [ secsquezcatlike ] , we introduce the class od squeezed cat - like states , we apply the optimization procedure , and we compare results with those obtained in the previous section . in section [ secnongaussmixed ] we will analyze the effect of noise on the performance of non - gaussian resources for cv teleportation . finally , in section [ conclusions ] we will draw our conclusions and discuss some perspectives for future research .

first we show calorimetric data obtained by the conventional dsc with a heating rate of 1/12 k s@xmath7 in fig . [ fig : figure1]a ( see methods ) . the grey curve in the top panel is a dsc curve for liquid 1 , which is obtained without annealing after liquid 1 is vitrified into its glassy state , glass 1 . there we can see the glass 1-to - liquid 1 transition , whose onset is located around 204 k. the signal also has a large exothermic ( heat releasing ) peak due to crystallization around @xmath8 k. on the other hand , the blue curve in the bottom panel shows a dsc heating curve for the glacial phase , which is prepared by annealing tpp for 600 min at 216 k until the transition is completed . we can see a change suggestive of the glass 2-to - liquid 2 transition , which starts around 210 k upon heating but is immediately followed by reverse llt from liquid 2 to liquid 1 and crystallization ( see below ) . the glass 2-to - liquid 2 transition is broader than the glass 1-to - liquid 1 transition , suggesting liquid 2 is less fragile than liquid 1 @xcite . ) and fast ( 10@xmath9 k s@xmath7 ) rates . * * a , * the results for the slow heating rate . the grey curve is a heating curve of liquid 1 without annealing and the blue curve is a heating curve of liquid 2 obtained after annealing for 600 min at 216 k. after the complete transformation from liquid 1 to the glacial phase ( glass 2 ) by constant - temperature annealing , the glass 1-to - liquid 1 transition signal completely disappears and instead there appears an endothermic peak at higher temperature , which is then followed by the significant exothermic peak due to crystallization . this exothermic peak makes it difficult to clarify the origin of the endothermic process . * b , * the results of flash dsc measurements . the black curve is obtained for a sample without annealing ( liquid 1 ) and the blue curve is for a sample after annealing ( the glacial phase , or glass 2 ) . the yellow dashed curve is taken after re - cooled from a point @xmath10 in the endothermic peak ( see the inset for the temperature protocol ) . the glass transition signal of liquid 1 is observed in the yellow dashed curve around 220 k , indicating that the glacial phase ( glass 2 ) has already returned to liquid 1 during the endothermic process before reaching @xmath10 . the grey curve is for a sample fully crystallized . , width=340 ] next we show dsc results obtained by the flash dsc with a heating rate of 10@xmath9 k s@xmath7 in fig . [ fig : figure1]b ( see methods ) . the black curve is a dsc curve for liquid 1 , which was obtained without annealing immediately after liquid 1 is vitrified into its glassy state , glass 1 , as in the top panel of fig . [ fig : figure1]a . @xmath4 observed with a heating rate of 10@xmath9 k s@xmath7 ( @xmath11 k ) is significantly higher than that observed with a slower heating rate of 1/12 k s@xmath7 ( @xmath3 k ) by the conventional dsc ( compare the black curve in fig . [ fig : figure1]b with the grey curve in the top panel of fig . [ fig : figure1]a ) . this is consistent with the general rule that the glass transition temperature increases with an increase in the heating rate . as shown in fig . [ fig : figure1]b ( see the black curve ) , the dsc signal obtained with the ultra - high speed heating exhibits no exothermic heat due to crystallization during the heating process , unlike the case of the slow heating in fig . [ fig : figure1]a ( see the grey curve ) . accordingly , there is no signature of crystal melting for a non - annealed sample ( the black curve ) in fig . [ fig : figure1]b , which is supposed to occur around 300 k ( see the grey curve in fig . [ fig : figure1]a ) . this clearly shows that the ultra - high speed heating successfully avoids the occurrence of crystallization after the glass 1-to - liquid 1 transition . after tpp is annealed at 216 k for 600 min , on the other hand , the glass transition signal of liquid 1 completely disappears and instead a large endothermic peak appears around 250 k ( see the blue curve in fig . [ fig : figure1]b ) upon heating . this indicates that there is no liquid 1 ( or glass 1 ) left in the glacial phase . this fact can not be explained by the nano - crystal scenario , since it assumes that the glacial phase is a mixture of glass 1 and nano - crystals ( see supplementary note 1 and supplementary figures 1 , 2 , 3 , 4 , 5 , and 6 for further evidence against the nano - crystal scenario ) . thus , we assign the glacial phase obtained by annealing to be glass 2 . in order to clarify what is happening at the endothermic peak , we employ the following special temperature protocol ( see the inset in fig . [ fig : figure1]b and the bottom part of panel a of fig . [ fig : figure2 ] ) . first we anneal a sample at 216 k for 600 min , which completely transforms liquid 1 to the glacial phase , and then quench it to @xmath12 k below @xmath2 . next we heat the system until the temperature @xmath10 indicated by the yellow point on the blue curve , keep it at @xmath10 for a period of 0.1 s , and then cool it again from @xmath10 to @xmath13 = 173 k below @xmath2 . here we use the cooling and heating rate of 10@xmath9 k s@xmath7 . the second heating from @xmath13 provides the yellow dashed dsc curve in fig . [ fig : figure1]b . we can clearly see the glass 1-to - liquid 1 transition signal in the yellow dashed curve . furthermore , the perfect overlap of the glass transition signal between the black curve and the yellow dashed curve in fig . [ fig : figure1]b suggests that the glacial phase ( or , glass 2 ) fully returns back to liquid 1 already much before crystal melting takes place in the heating process ( more specifically , either before reaching @xmath10 in the first heating process or during 0.1 s kept at @xmath10 ) . thus , the endothermic peak around 250 k in the blue curve should not be associated with the crystals that should melt around 300 k. ( top ) and for reverse llt as a function of the re - cooling temperature @xmath10 ( bottom ) . the experimental protocols for them are also show above and below the picture . * b , * the forward llt process from liquid 1 to liquid 2 , probed by the reverse llt process . the annealing time ( @xmath14-)dependence at @xmath15=216 k. the numbers for lines with various colours denote @xmath14 . * c , * the reverse llt process from liquid 2 to liquid 1 . the black dotted curve is for the sample without annealing and the black solid curve is for the sample annealed for 600 min at 216 k. the other curves are obtained for the second heating ( see the protocol ) . the numbers for lines of various colours denote @xmath10 . * d , * the annealing time ( @xmath14-)dependence of the heat released by the reverse llt upon heating for a sample annealed at 216 k. the inset illustrates how to estimate the transition heat by integration . * e , * the re - cooling temperature ( @xmath10- ) dependence of the heat released upon heating for a sample annealed at 216 k for 600 min , in which llt is completed . we can see that the onset of the reverse llt , at which the endothermic heat starts to decrease , is located around 235 k. this temperature is located slightly above @xmath16 . , width=377 ] we can see that the crystal melting behaviour around 300 k is almost perfectly the same between the blue curve and the yellow dashed curve ( see fig . [ fig : figure1]b ) . this result clearly indicates that the crystals are formed exclusively during annealing and they are not affected by heating and cooling below @xmath10 , suggesting that liquid 2 is prone to crystallization compared to liquid 1 . we also note that the amount of the heat of fusion in the blue curve is much smaller than that in the grey curve , which is for a fully crystallized sample . this can be explained as follows : crystallization takes place preferentially in liquid 2/glass 2 domains , which are newly formed during annealing , but its glassy nature inhibits both nucleation and growth of crystals . we show experimental results on the forward and reverse llt processes in much more detail ( see the top part of fig . [ fig : figure2]a for the protocol and the resulting phase change process as a function of @xmath14 ) . the annealing time @xmath14-dependence of the first heating curve is shown in fig . [ fig : figure2]b . the glass 1-to - liquid 1 glass transition signal becomes smaller with an increase in @xmath14 and completely disappears for @xmath17 min , indicating that the liquid 1-to - glass 2 transition is completed around this annealing time . on the other hand , a new endothermic peak appears for @xmath18 min and continues to grow with an increase in @xmath14 . another important fact is that after the endothermic peak the heat capacity of the liquid is the same as that of liquid 1 , which can be seen from the fact that above 260 k all curves almost coincide with each other with the curve of @xmath19 for liquid 1 . this clearly indicates that the endothermic peak is associated with the transition from glass 2 to liquid 1 . furthermore , we can see in fig . [ fig : figure2]b that the endothermic peak position shifts to a higher temperature with an increase in @xmath14 . figure [ fig : figure2]d shows the @xmath14-dependence of the total heat released by reverse llt , which should be proportional to the amount of liquid 2 formed during @xmath14 . we also show the @xmath10-dependence of the second heating curve in fig . [ fig : figure2]c ( see the bottom part of fig . [ fig : figure2]a for the protocol and the schematic figure showing the resulting phase change process as a function of @xmath10 ) . here the sample is kept for 0.1 s at @xmath10 before re - cooling from @xmath10 . with an increase in @xmath10 , the endothermic peak becomes smaller and the glass 1-to - liquid 1 transition signal emerges and gradually becomes larger . figure [ fig : figure2]e shows the @xmath10-dependence of the heat released during the reverse llt , which should be proportional to the amount of liquid 2 remaining after heated to @xmath10 . on the basis of these results , we discuss the origin of the new endothermic peak emerging after annealing at 216 k ( see the blue curve in fig . [ fig : figure1]b ) . there are the following three possible origins for the endothermic peak appearing around 250 k : ( i ) the glass 1-to - liquid 1 transition , ( ii ) the glass 2-to - liquid 2 transition , and ( iii ) the reverse llt from liquid 2 to liquid 1 . whichever the transformed state contains liquid 1 or liquid 2 , the system is initially in a glassy state and , thus , it should exhibit a glass transition signal upon heating before finally returning to liquid 1 @xcite . there is a difference in the heat flow level between before and after the endothermic peak , indicating the difference in the heat capacity @xmath20 ( see the blue curve in fig . [ fig : figure1]b ) . this is consistent with the occurrence of glass transition . however , we show below that the glass transition can not be a primary origin of the endothermic peak . firstly we consider possibility ( i ) that the endothermic peak is due to the glass 1-to - liquid 1 transition . although the position of the endothermic peak is significantly different from the glass transition peak of liquid 1 ( @xmath14 = 0 ) , it alone does not immediately mean that the system is liquid 2 and not liquid 1 . this is because ageing can generally shift the glass transition peak towards a higher temperature . thus , even if the peak is due to the glass 1-to - liquid 1 transition , the peak position can depend on @xmath14 : the ageing of glass 1 should continuously shift the peak towards a higher temperature and increase the magnitude of the glass transition signal . contrary to this expectation , however , fig . [ fig : figure2]b tells us that an increase in @xmath14 reduces the signal of the glass - to - liquid transition and leads to the emergence of a new endothermic peak at a much higher temperature and the increase of its hight . this observation indicates that there are clearly two transitions with different origins . thus , the presence of the two distinct transitions can not be explained by scenario ( i ) based on the ageing of glass 1 . this conclusion is also supported by the fact that after the transition ( @xmath21 min ) the glass transition signal associated with liquid 1 component completely disappears . next we consider possibility ( ii ) that the endothermic peak is mainly due to the glass 2-to - liquid 2 transition . the step - like change at a lower temperature is definitely associated with the glass 1-to - liquid 1 transition at least for a rather short annealing time @xmath14 . on the other hand , the endothermic peak appearing after annealing should be associated with liquid 2 formed during annealing . we note that the temperature shift of the peak towards a high temperature with an increase in @xmath14 does not stop even after the transition is completed . however , the total heat released during the transition becomes constant after the completion of the transition ( @xmath22 min ) , as shown in fig . [ fig : figure2]d . the ageing of a glass should lead to the simultaneous increase in both the transition temperature and peak area . the lack of this feature indicates that the heat involved in the endothermic peak can not be explained by the glass 2-to - liquid 2 transition alone , even taking the ageing effect into account . furthermore , as shown in fig . [ fig : figure1]b , glass 2 has already returned to liquid 1 at @xmath10 upon heating , the endothermic peak should involve the reverse llt ( see also below for the further supporting evidence ) . thus we conclude that the endothermic peak is primarily not due to the glass 2-to - liquid 2 transition , although it should contribute partially . finally , we consider the remaining possibility ( iii ) that the endothermic peak should come mainly from the reverse llt from liquid 2 to liquid 1 . this scenario is strongly supported not only by the above - mentioned transformation of glass 2 to liquid 1 before reaching @xmath10 but also by the fact that the heat released by the reverse llt ( @xmath23 j / g ) ( see fig . [ fig : figure2]d , e ) is comparable to the heat absorbed by the forward llt ( @xmath24 25 - 27 j / g ) . here we note that the transition heat of the reverse llt does not depend on the heating rate in the range of 500 @xmath24 2000 k s@xmath7 and is almost constant within @xmath25 1 % . this reflects that the transition is between the well - defined glass 2 state , which is almost uniquely determined by the annealing temperature and the annealing time , and the liquid 1 state . the difference in the transition heat between the reverse and forward llt may come from the contribution of the glass 2-to - liquid 2 transition . the onset of the glass 2-to - liquid 2 transition marks the onset of the heat release ( see below on the details of the glass transition behaviour ) . we note that the glass 2-to - liquid 2 transition provides the system with mobility , which is necessary for the reverse llt to proceed . in this reverse llt scenario , we can explain the shift of the peak position towards a higher temperature with an increase in @xmath14 as a consequence of the ageing of glass 2 : glass 2 becomes more stable and its glass transition temperature becomes higher with @xmath14 , leading to the shift of the onset of the reverse llt towards a higher temperature . during annealing at @xmath15 , a system gradually transforms from liquid 1 to the glass state of liquid 2 ( glass 2 ) with @xmath14 . reflecting this , the total heat released during the reverse llt should increase with an increase of @xmath14 . for @xmath22 min , however , it becomes constant since llt is completed , i.e. , the system almost perfectly becomes glass 2 ( see fig . [ fig : figure2]b , d ) . this saturation indicates that the order parameter @xmath6 in glass 2 becomes almost constant for @xmath22 min . so we conclude that the transition behaviour shown in fig . [ fig : figure2]b consists of the step - like glass 2-to - liquid 2 transition and the endothermic peak due to the reverse llt from liquid 2 to liquid 1 ( see below on the separation of the two transitions ) . next we consider the experimental results using the special temperature protocol ( see the bottom part of fig . [ fig : figure2]a for the protocol and fig . [ fig : figure2]c for the results ) . in the above , we show that liquid 2 already fully returns to liquid 1 at the yellow point marked on the blue curve in fig . [ fig : figure1]b . particularly , fig . [ fig : figure2]c shows that the glass transition signal associated with liquid 1 gradually recovers during the endothermic process with increasing @xmath10 , clearly supporting the above - mentioned scenario that this endothermic peak is due to the reverse llt process ( liquid 2 @xmath26 liquid 1 ) . this is also consistent with the fact that the reverse llt process from liquid 2 to liquid 1 should be an endothermic process since the forward llt from liquid 1 to liquid 2 during isothermal annealing is an exothermic process @xcite . we indeed confirm the amount of heat associated with the transition is about the same between the forward and reverse processes , as mentioned above . we can see in fig . [ fig : figure2]c that the peak position slightly shifts towards a higher temperature for higher @xmath10 . this suggests that more stable parts of liquid 2 with higher @xmath6 transform to liquid 1 at higher @xmath10 . now we focus on the glass transition behaviour taking place prior to the reverse llt upon heating ( see also supplementary note 2 and supplementary figures 7 and 8) . the glass transition behaviours shown in fig . [ fig : figure2]b and c suggest that , in both processes of the forward and reverse llt , the two disordered phases of liquid 1 and 2 coexist and transform reversibly with each other ( see the schematic pictures in fig . [ fig : figure2]a ) . this is the first unambiguous evidence not only for the reversibility and the first - order nature of llt in molecular liquids but also for the fact that the transition is between two distinct liquid states . the presence of the two glass transitions and the direct reversibility of the transition between the two liquid phases can be naturally explained by the llt scenario , but not by the nano - crystal scenario . to be more quantitative , we identify the onset of the glass - to - liquid transition of the glass state obtained by various annealing time @xmath14 , as shown in fig . [ fig : figure3 ] ( see the inset of panel e on the determination of the onset of the glass transition ) . the glass transition behaviour provides crucial information on the type of llt , i.e. , whether llt is nucleation - growth ( ng)-type or spinodal decomposition ( sd)-type @xcite . we note that ng - type llt proceeds in a metastable state while accompanying nucleation of liquid 2 droplets in liquid 1 , whereas sd - type llt proceeds in an unstable state by a continuous transformation of liquid 1 to liquid 2 . k and 216 k , respectively . * c - e , * the same observed during sd - type llt observed at @xmath27 , 210 , and 208 k , respectively . * f , * the dependence of the glass transition behaviour as a function of @xmath10 in the reverse llt process . the open circles show the onset temperatures of glass transition and the widths of the half transparent belts roughly represent possible errors in their determinations . an example of the estimation of the onset temperature of a glass transition is shown in the inset of panel e. the arrows indicate the directions of the increase in @xmath14 for panels a - e and that in @xmath10 for panel f. , width=529 ] when llt proceeds above @xmath28 k , we have two sequential glass transitions upon heating . for @xmath15=216 k , for example , we identify @xmath29 k and @xmath30 k ( see fig . [ fig : figure3]b ) . interestingly , the positions of the onsets of the two glass transitions do not depend upon the annealing time within errors ( @xmath31 k ) . this is a characteristic feature of ng - type of llt , reflecting that , for ng - type llt , liquid 2 with the final order parameter value is nucleated in liquid 1 with the initial order parameter value and thus the order parameter changes discontinuously from that of liquid 1 to that of liquid 2 . this coexistence of the two ` liquid ' phases during transformation can be regarded as a direct manifestation of llt and its first - order nature @xcite . here it is worth mentioning that the glass 2-to - liquid 2 transition behaviour is not so clear compared to the glass 1-to - liquid 1 transition . the spinodal temperature , @xmath32 , or the stability limit of liquid 2 against liquid 1 upon heating , is estimated to be around 235 k , which is located only slightly above the onset of the glass 2-to - liquid 2 transition . thus , the glass transition does not complete at this temperature within a short time . this means that the plateau of @xmath20 after the glass 2-to - liquid 2 transition never appears , making it difficult to observe a typical glass transition signal . for the reverse llt to take place , the system needs to gain mobility . once the system starts to gain mobility due to the glass 2-to - liquid 2 transition , the reverse llt is immediately initiated . in other words , the glass transition and the reverse llt almost simultaneously take place , making clear separation between the glass 2-to - liquid 2 transition and the reverse llt intrinsically difficult . when llt proceeds below @xmath33 ( @xmath34 k ) , on the other hand , we can see that there is only one glass transition , whose onset temperature continuously and gradually shifts from that of glass 1 to that of glass 2 , as shown in fig . [ fig : figure3]c - e . we note that we analyse the data only after the ageing is completed ( see supplementary figure 9 and supplementary note 2 ) . this glass transition behaviour is a characteristic feature of sd - type llt , where the order parameter changes continuously with time @xcite . unlike the ng - type llt , the reverse llt starts immediately after the first glass transition step of the glass state , whose transition temperature is located between those of glass 1 and 2 in the process of llt . here it may be worth explaining why we may conclude that there is only one glass transition : below @xmath33 ( @xmath34 k ) , the @xmath20 exhibits a minimum following the glass - transition peak , but its value is larger than the heat capacity of liquid 1 ( see the curve for @xmath35 min at @xmath15=210 k in fig . [ fig : figure3]d and the curves for @xmath36 min at 208 k in fig . [ fig : figure3]e ) . considering that liquid 2 is in a more ordered state than liquid 1 , the heat capacity of liquid 2 is expected to be smaller than that of liquid 1 . thus , the @xmath20 minimum larger than that of liquid 1 should stem from an additional contribution to @xmath20 from the reverse llt . this means that the reverse llt already starts before the completion of the glass 2-to - liquid 2 transition and there is no other glass transition . we note that such behaviour is never observed for the case of ng - type llt ( see fig . [ fig : figure3]a and b ) . this simultaneous occurrence of the glass transition and the reverse llt is further supported by a clear two - step feature in the dsc curves for @xmath37 min at @xmath15=208 k in fig . [ fig : figure3]e . the first step is the glass - to - liquid transition and the second one is the reverse llt , unambiguously indicating that there are two distinct sequential transitions . finally , we stress that these glass transition behaviours can not be explained by the nano - crystal scenario , in which there should be only one glass transition from glass 1 to liquid 1 at @xmath4 . , whereas for sd - type llt there is only one glass transition and its onset temperature continuously increases with @xmath14 . , width=377 ] we compile all the data of the onset of the glass transition temperature in fig . [ fig : figure4 ] . there we can clearly see that for sd - type llt below @xmath38 k the onset temperature of the glass transition gradually and continuously shifts towards a high temperature with @xmath14 whereas for ng - type llt above @xmath33 those of glass 1 and 2 stay almost constant as a function of @xmath14 until the completion of the transformation , whose timing is indicated by the arrows in fig . [ fig : figure4 ] . this is fully consistent with the phenomenology of ng- and sd - type phase transformation @xcite . we stress that the above glass - transition behaviours are fully consistent with the characteristics of pattern evolution during llt revealed by optical microscopy observation @xcite . the microscopy observation suffered from a criticism stemming from a resolution problem : people may suspect that continuous evolution of smooth density fluctuations observed for sd - type llt may be merely a consequence of that droplets are actually formed but too small to be optically resolved . our dsc results clearly indicate that this is not the case and sd - type llt of the continuous nature indeed takes place below @xmath33 . thus , our finding strongly supports the physical picture of the two - order - parameter model of llt @xcite . finally , in the reverse llt process , we can see that the onset temperature of the glass transition continuously shifts from that of liquid 2 to that of liquid 1 , as shown in fig . [ fig : figure3]f . this indicates that the reverse llt takes place via sd - type transformation . this is consistent with the fact that the process takes place almost immediately ( less than 0.1 s ) without an incubation time . this fast transformation process implies that the transformation takes place in a liquid state with fast dynamics , i.e. , far above the glass transition from glass 2 to liquid 2 , @xmath16 . the onset temperature below which the transition heat starts to decrease is located around @xmath39 k ( see fig . [ fig : figure2]e ) . thus , this @xmath10 marks the stability limit of liquid 2 against liquid 1 upon heating , i.e. , the temperature above which liquid 2 becomes unstable against liquid 1 , i.e. , @xmath40 k. at ambient pressure , there is a rather large difference between @xmath33 and @xmath32 , but this difference is expected to decrease with an increase in pressure and should disappear at the critical pressure @xmath41 . there the two spinodal temperatures @xmath33 and @xmath32 should merge to the critical temperature @xmath42 , which should be located above 235 k. from this , we can conclude that @xmath32 is located around 235 k for @xmath43 k. this can also be confirmed in fig . [ fig : figure3]e ( see the curve at @xmath44 min ) , where we can see the glass transition and reverse llt separately . we note that it is located slightly above @xmath16 ( @xmath45 k for @xmath43 k at a heating rate of @xmath46 k s@xmath7 ) . in summary , we reveal by ultra high - speed calorimetry that , upon heating , glass 2 that is formed by annealing liquid 1 at a low temperature for a long time , first transforms into liquid 2 via the glass transition and then almost simultaneously liquid 2 becomes liquid 1 via the reverse llt . our study not only shows the reversibility of llt of tpp , but also its first - order nature from the coexistence of the two distinct liquid phases during llt and the transformability between them . furthermore , the glass transition behaviours of an intermediate state formed during llt upon fast heating tell us that there are two types of llt , ng - type and sd - type llt , which are typical non - equilibrium dynamical processes of the first - order phase transition respectively in its metastable and unstable state . we successfully reveal the discontinuous and continuous nature of the order parameter evolution for ng - type and sd - type llt respectively from the glass transition behaviour of a transient state during llt . this firm experimental confirmation of the first - order liquid - liquid transition in a single - component molecular liquid may initiate a new phase of theoretical and experimental research on the physical nature of this intriguing phase transition phenomenon and lead to a deeper understanding of the liquid state of matter . it may also contribute to the resolution of the controversy on llts of various systems such as water , which are also supposed to occur in a non - equilibrium metastable state @xcite : our experimental method may be useful for isolating llt from nano - crystallization for other systems . * material . * triphenyl phosphite ( tpp , 99.7% purity ) was purchased from across organics and used it without further purification . the melting point @xmath47 = 297 k , whereas the glass transition temperature of liquid 1 @xmath4 = 204 k at a heating rate of 5 k min@xmath7 . * calorimetry measurements . * we used an ultra high - speed dsc ( mettler - toledo flash dsc 1 ) and a conventional dsc ( mettler - toledo dsc 1 ) . in ultra high - speed dsc measurements , the sample mass was 20 - 50 ng , which was estimated for each sample by comparing the heat of fusion of a fully crystallized sample obtained by the flash dsc with that obtained by the conventional dsc . the fully crystallized sample was obtained by heating at 20 k min@xmath7 from 173 k. in the forward and reverse llt experiments , we used the protocol shown in fig . [ fig : figure2]a , where the cooling and heating rate were 1000 k s@xmath7 , the lowest temperature was 173 k , a waiting time of 0.1 s was inserted between each scan for stabilization of the instrument . in conventional dsc measurements , the sample mass was 11.63 mg , the cooling and heating rates were 10 k min@xmath7 and 5 k min@xmath7 respectively , and the lowest temperature was 173 k. all calorimetric measurements were performed under the n@xmath48 atmosphere . * acknowledgements * the authors thank c.a . angell for valuable discussions . this study was partly supported by specially promoted research ( 25000002 ) , scientific research ( s ) ( 21224011 ) , and scientific research ( c ) ( 16k05510 ) from the japan society for the promotion of science ( jsps ) . s upon heating at 1000 k / s . * endothermic peaks around 240 - 270 k and melting peaks of bulk crystals around 300 k are separated , suggesting that the origin of endothermic peak is not due to the melting of nano - crystals ( see text ) . , width=453 ] s , upon heating at 1000 k / s . * * a , * @xmath49-dependences of the transition heat of the reverse llt , the exothermic signal appearing above @xmath50 k due to crystallization , and the melting of crystal . * b , * @xmath49-dependences of the peak temperatures for the reverse llt and melting . , width=302 ] k. * * a , * reverse llt processes upon second heating of samples , which are formed by annealing tpp at @xmath15=225 k for 200 min , then heated to @xmath10 , kept for 0.1 s there , and rapidly cooled to a low temperature . the temperature protocol and @xmath10 are shown in the inset . note that only a part of the system returns to liquid 1 due to the crystallization . * b , * @xmath51-dependence of the heat released upon heating estimated from the results in panel a. the inset explains how to estimate the total heat released . the onset of the reverse llt is found to be located around 242 k. , width=340 ] during llt for various @xmath15 s . * we calculate @xmath52 by integrating the area of each heat process . we use a signal level at @xmath53 = 0 as the base line of the integration and an example of our analysis is shown in the first top left panel . , width=566 ] dependence of raman spectra . * we use the polarized incident laser light ( 532 nm ) for excitation and detect the scattered light of all polarizations . all measurements are made after llt is completed . the signals above 225 k indicates the presence of crystals , whereas those below 223 k are typical amorphous raman spectra and do not show any indication of the presence of crystals . , width=340 ] . the @xmath4 ( blue circles ) and @xmath16 ( red circles ) do not change as a function of @xmath14 above 214 k ( ng - type ) , whereas @xmath2 ( blue circles ) continuously changes below 214 k ( sd - type ) . the inset in the panel of 212 k shows how to determine @xmath54 . , width=491 ] -dependence of the glass transition behaviours during the reverse llt . * in this case , we also see the gradual continuous change of @xmath2 during the reverse llt transition , which is suggestive of the sd - type nature of the transformation . , width=340 ] there have been long - standing arguments that the transition we focus here is nano - crystal formation @xcite . in this scenario , the endothermic peak observed upon heating of the glacial phase should be due to the melting of nano - crystals . although the endothermic peak is located at a temperature much lower than the melting point of bulk crystal , this might be explained by very small sizes of nano - crystals and their defective structures @xcite . if we assume this , the system can have two melting temperatures of crystals having very different sizes but with the same structure . this is a possible interpretation of the phenomenon . it is rather difficult to deny this scenario in a clear manner since nano - crystals are formed in a supercooed metastable liquid state anyway . this is the source of the long - lasting serious controversies on the nature of the glacial phase @xcite . in the main text , we describe a few reasons why this scenario is difficult to explain our observation . here we further discuss this possibility and show convincing experimental evidence that this transition can not be explained by nano - crystal formation but should be llt . if we try to explain the endothermic peak by the nano - crystal scenario , we need to assign this peak as ( i ) the melting of very small crystals , ( ii ) the melting of a new type of crystal distinct from the known crystal , or ( iii ) the solid - state transition from an ordered to a disordered crystal . first of all , there is no sign of the glass transition associated with liquid 1 after the formation of the glacial phase ( see fig . 1b in the main text ) . in the nano - crystal scenario , the glacial phase is considered as a mixture of glass 1 and nano - crystals . thus , the absence of the glass transition of liquid 1 means that the system is filled with nano - crystals without any amorphous parts . this is not consistent with a much smaller heat of fusion of the glacial phase than crystals ( see fig . 1b in the main text ) and previous x - ray scattering experimental results @xcite . however , one may still argue that crystals are so disordered and thus the heat of fusion is much smaller than that of good crystals ( see below ) . the melting of such small unstable nano - crystals might happen at a much lower temperature than @xmath55 @xcite . then the endothermic peak might be the melting of such extremely small defective crystals . in this scenario , however , the gradual change of the glass transition temperature as a function of @xmath10 can not be explained since there should be only liquid 1 that contributes to the glass transition . furthermore , according to the x - ray scattering measurements , there is any indication of neither the formation of a new type of crystals , nor the disappearance or position change of the bragg peaks around 250 k. all these indicate that the above - mentioned scenarios based on nano - crystals can not explain our dsc results . next we show the annealing temperature @xmath49-dependence of the heating curve of tpp samples after annealed for the same fixed duration of 600 min , in supplementary figure [ fig : tw600min ] . note that the transition is completed above @xmath49 = 216 k before 600 min , whereas it is not completed below 216 k. here we focus on the behaviours of the second transition appearing after the glass transition and the third transition around 300 k , which is the melting of crystals . we can see that the endothermic peak position of the second transition shifts towards a higher temperature with an increase in @xmath49 . the peak eventually disappears for @xmath56 k and the highest observable peak temperature is located around 270 k. on the other hand , the melting of bulk crystals always takes place around 300 k irrespective of @xmath49 . we summarize the heat associated with phase transitions ( the reverse llt , crystallization during heating , and the crystal melting ) in supplementary figure [ fig : tw600result]a and the peak temperatures of the reverse llt and the crystal melting in supplementary figure [ fig : tw600result]b . the peak temperature of the endothermic peak of the reverse llt monotonically increases with an increase in @xmath15 . below @xmath57 k , the effect may be intrinsic , but above 220 k the peak temperature shift might be due to the presence of nano - crystals embedded in glass 2 . the former is because there are few nano - crystals formed during annealing for @xmath58 k. on the other hand , the latter is because the amount of nano - crystals formed during the heating , which can be estimated by subtracting the heat released upon crystallization from the heat absorbed upon melting ( see supplementary figure 2a ) , starts to increase above 220 k with an increase in @xmath15 the increase in the onset temperature of the glass transition of liquid 2 and the reverse llt for @xmath59 k may be due to the effects of ageing , since the ageing of glass 2 proceeds more quickly at a higher @xmath15 . in relation to this origin of the temperature shift , it should be noted that according to our previous x - ray scattering studies , the number density of locally favoured structures , or the order parameter @xmath6 , does not depend on the annealing temperature @xmath15 . since the @xmath15-dependence of the endothermic position is not fully understood , however , we need further investigation to clarify its origin . if we assume that the endothermic peak is due to the nano - crystal melting , the shift of the peak position towards a higher temperature can be explained by the larger size of crystals and/or their higher perfectness for higher @xmath49 . this scenario suggests that the melting peak of nano - crystals should continuously shift towards that of bulk crystal . however , our data shows that the endothermic peak can not exist above 270 k , indicating its discontinuous jump to the melting peak of bulk crystal between @xmath60 k and 237 k ( see supplementary figure 1 ) . this discontinuity between the endothermic peak and the melting peak of bulk crystals suggests that the endothermic peak is not due to nano - crystal melting . we also note that the fact that the endothermic peak due to the reverse llt disappears above 232 k indicates that only below this temperature llt takes place . thus , the binodal temperature of llt is determined as @xmath61232 k. this value of @xmath62 is consistent with our previous estimation , @xmath63 k @xcite . if we anneal a sample above this temperature , we have only crystallization phenomena and llt can not be induced . the results shown in supplementary figure [ fig : tw600min ] also show clearly that the onset temperature of the reverse llt increases with an increase in @xmath15 , thus suggesting the increase of @xmath16 with @xmath15 . we confirm that this is indeed the case . supplementary figure [ fig : trc]a shows the @xmath10-dependence of the endothermic signal coming from the reverse llt for @xmath15=225 k. supplementary figure [ fig : trc]b plots the @xmath51-dependence of the heat released upon heating estimated from the results in supplementary figure [ fig : trc]a . from this , we can estimate the onset of the reverse llt to be located around 242 k , suggesting that @xmath64 @xmath65 242 k for glass 2 formed at @xmath15=225 k. in the main text we show that @xmath64 235 k for @xmath66 k. thus , this result clearly indicates that @xmath16 and @xmath32 increase with an increase in @xmath15 . next we show in supplementary figure [ fig:225k ] the time dependence of the dsc heating curve for tpp samples annealed at 225 k. at this temperature , the melting peak of bulk crystals around 300 k is very large compared to those for lower @xmath15 and the amount of the crystals increases with an increase in the annealing time . on the other hand , the exothermic heat between the endothermic peak and the melting peak upon heating does not increase so much by increasing the annealing time , clearly indicating that the crystals which melt around 300 k is formed during llt and not during heating . however , it is not reasonable to assume that nano - crystals and bulk crystals are formed simultaneously in the annealing process . nano - crystals can be formed only under the situation that crystals can not grow after nucleation . extremely low mobility or internal frustration makes such a situation possible , but should make the formation of bulk crystals impossible . thus , we conclude that it is difficult to have nano - crystals and bulk crystals at the same time in an ordinary situation . finally , we evaluate the transformation heat of each process : the endothermic peak due to the reverse llt , the exothermic broad part due to crystallization , and the crystal melting peak at several annealing temperatures ( see supplementary figure [ fig : deltae ] ) . the time evolution of the endothermic heat reflects the process of the endothermic transition . it should be noted that the melting component grows before the endothermic peak grows . this trend is evident particularly at @xmath15 below 216 k. for example , at @xmath67 k , the melting component appears around the annealing time of 30 - 60 min , whereas the endothermic peak appears around 1000 min . this fact strongly suggests that the endothermic transition and crystallization are the processes of essentially different nature . this also indicates the difficulty of the nano - crystal scenario and supports the llt scenario . in relation to the above , it should be noted that the crystal formation before the transition has been observed by neither light scattering nor x - ray scattering measurements . we show in supplementary figure [ fig : raman ] raman scattering data at various @xmath49 s , which are taken after llt is completed . there are small peaks due to crystals in the data above @xmath49 = 225 k , whereas such peaks are absent below 223 k and the spectra are typical amorphous signals , indicating the absence of crystals . these results indicate that crystals detected by our dsc in the process of the transformation during annealing should be unusually small , i.e. , nano - crystals . according to our optical microscopy observation @xcite , there are two types of the dynamic processes in llt of tpp : nucleation - growth ( ng)-type and spinodal - decomposition ( sd)-type llt . this classification is based on the types of pattern evolution . we confirm the presence of these two types of llt from the glass - transition behaviour of tpp during the llt process , as discussed in the main text . we stress that the glass transition behaviour is specific to a liquid state . the existence of the two glass transition temperatures for a single - component liquid and the presence of two types ( ng - type and sd - type ) temporal changes of the glass transition temperatures strongly support that the transition is indeed llt and the endothermic peak observed on heating is the reverse llt . each feature of ng - type and sd - type llt appears in the behaviour of the glass transition temperature , @xmath2 . supplementary figure [ fig : onsettg1-all ] is the annealing time dependence of the glass transition temperature for several annealing temperatures . we estimate @xmath2 as the onset temperature of the glass transition upon heating . the method to estimate @xmath54 is shown in the inset of the panel of @xmath27 k. the @xmath2 determined in this way is shown by an open circle on each curve . we analyse only the data after ageing of liquid 1 to see the change of the glass transition temperature due to llt alone . above 216 k , @xmath54 shows almost no change with the progress of llt , indicating that this glass transition is always that of liquid 1 . the magnitude of the glass transition step , or the amount of liquid 1 , monotonically decreases with the annealing time , reflecting the nucleation and growth of liquid 2 domains in the sample , which was observed with optical microscopy @xcite . this is the typical behaviour expected for the ng - type llt . on the other hand , @xmath54 continuously shifts towards a higher temperature with the waiting time for @xmath15 below 212 k , suggesting that the transition from liquid 1 to liquid 2 is continuous . this behaviour is consistent with sd - type llt . these results clearly indicate that the dynamical process of llt can be classified into ng - type and sd - type ( see also fig . 4 in the main text ) . this is fully consistent with our microscopic observation of pattern evolution @xcite , strongly supporting the llt scenario @xcite . supplementary figure [ fig : reverse ] shows the @xmath10-dependence of the glass transition behaviours during the reverse llt . the gradual continuous change of @xmath2 during the reverse llt transition suggests that this process may be sd - type transformation , which is consistent with its rather rapid transformation . finally we show the effects of ageing on the onset temperature of the glass transition by taking the data of @xmath67 k as an example . supplementary figure [ fig : ageing]a shows the dsc curves in the glass transition region for @xmath67 k , whereas supplementary figure [ fig : ageing]b plots the estimated onset temperature against the annealing time @xmath14 . we can see that the initial increase of @xmath2 is due to the ageing of glass 1 . the time region of the constant @xmath2 of glass 1 indicates that the ageing of glass 1 is completed after 10@xmath7 min . the final continuous increase after 10@xmath9 min is due to sd - type llt , which can be seen in supplementary figure [ fig : ageing]b .

* liquid - liquid transition is an intriguing phenomenon in which a liquid transforms into another liquid via the first - order transition . for molecular liquids , however , it always takes place in a supercooled liquid state metastable against crystallization , which has led to a number of serious debates concerning its origin : liquid - liquid transition vs. unusual nano - crystal formation . thus , there have so far been no single example free from such debates . here we show the first firm experimental evidence that the transition is truly liquid - liquid transition and not nano - crystallization for a molecular liquid , triphenyl phosphite . we kinetically isolate the reverse liquid - liquid transition from glass transition and crystallization with an extremely high heating rate of flash differential scanning calorimetry , and prove the reversibility and first - order nature of liquid - liquid transition . our finding not only deepens our physical understanding of liquid - liquid transition but also will initiate a new phase of its research from both fundamental and applications viewpoints . * even for a single - component substance , there can be more than two liquid states @xcite . the transition between these different liquid states is called `` liquid - liquid transition ( llt ) '' . llt is one of the most mysterious phenomena in liquid science and its presence and absence have often been debated for various systems . since this problem is of fundamental importance in our understanding of the liquid state , llt has kept attracting considerable attention . the presence of llt has been reported for both molecular systems ( water @xcite , triphenyl phosphite ( tpp ) @xcite , n - butanol @xcite , and possibly d - mannitol @xcite ) and atomic systems ( sulphur @xcite , phosphorus @xcite , silicon @xcite , germanium @xcite and @xmath0-@xmath1 @xcite ) . recently , llt was also reported for metallic glass - formers @xcite . however , none of these examples is free from controversy . the situation is more complicated for molecular liquids than for atomic liquids , since llt always takes place in a supercooled state metastable against crystallization for molecular systems @xcite . for atomic systems , on the other hand , the situation is better , since llt often takes place in an equilibrium liquid state : for example , it was shown that liquid p shows a first - order like transition from p4 tetrahedra to polymereric p chain structure @xcite and liquid s transforms into different polymeric structures upon heating @xcite . one of the hottest and long - standing debates is on the nature of an unconventional amorphous state called `` glacial phase '' discovered by kivelson and his coworkers @xcite for a molecular liquid , triphenyl phosphite ( tpp ) . we note that tpp is one of the most well - studied molecular systems which are expected to have llt . when tpp is kept at a low temperature near but still above the glass transition temperature @xmath2 ( @xmath3 k ) , a supercooled state of liquid 1 slowly transforms to an apparently amorphous state distinct from its ordinary glass state . since this transformation occurs above @xmath2 of liquid 1 , @xmath4 , the final amorphous state must not be a glass state of liquid 1 , glass 1 . we showed that this phenomenon can naturally be explained by llt from liquid 1 to a glass state of liquid 2 , glass 2 @xcite . in this scenario , thus , the glacial phase is glass 2 . in the following , for simplicity , we use the term `` llt '' to express the transition between liquid 1 and liquid 2/glass 2 without distinguishing whether liquid 2 is in a liquid or glass state . besides the llt scenario , many other explanations were also proposed for the nature of the glacial phase ; for example , the glacial phase was interpreted as a mixture of glass 1 and nano - crystals @xcite , a liquid crystal or plastic crystal @xcite , and an unconventional crystal called defect - ordered crystal @xcite . x - ray diffraction data of the glacial phase formed at a very low temperature show only broad amorphous peaks and no sharp bragg peaks , indicating the absence of distinct translational order in the glacial phase @xcite . this feature can not be explained by the plastic - crystal scenario . the glacial phase prepared at a very low temperature does not exhibit distinct birefringence @xcite , which can not be explained by the liquid - crystal scenario . on the other hand , the glacial phase formed at a rather high temperature exhibits not only weak birefringence @xcite but also small bragg peaks @xcite . this can naturally be explained by the presence of nano - crystals in the glacial phase . then the question is whether the glacial phase is primarily glass 2 or just a mixture of glass 1 and nano - crystals whose size decreases with a decrease in the annealing temperature , at which the glacial phase is formed . in the latter scenario , the absence of the bragg peaks in the glacial phase formed at a very low temperature is ascribed to an extremely small size of nano - crystals . such debates may also originate from the counter - intuitive impression about llt . according to classical liquid - state theory @xcite , the liquid state can be described by a single order parameter , density @xmath5 . provided that a liquid is in a random disordered state , it is hard to accept the presence of two liquids with different densities intuitively . however , once we accept that we need an additional scalar order parameter besides density to describe the state of a liquid , llt is no longer counter - intuitive and can be accepted naturally . on the basis of this idea and along the spirit of the pioneering works by strssler and kittel @xcite and rapoport @xcite , we proposed a two - order - parameter model of liquid - liquid transition @xcite . in this picture , for example , llt in atomic systems like sulphur and phosphorous @xcite can be explained by the distinct change in the locally favoured structures stabilized by chemical ( or , covalent ) bonding . similarly , locally favoured structures can also be formed by directional hydrogen bonding for molecular liquids . according to our two - order - parameter model @xcite , the order parameter governing llt is the fraction of locally favoured structures , @xmath6 , and then llt is regarded as a gas - liquid - like transition of the order parameter @xmath6 : liquid 1 is a gas - like state with low @xmath6 whereas liquid 2 is a liquid - like state with high @xmath6 . since locally favoured structures are created and annihilated independently , their number density is not conserved and thus @xmath6 is a non - conserved scalar order parameter . our recent x - ray scattering study @xcite revealed that upon llt of tpp locally favoured structures whose size is a few nm are formed and its number density monotonically increases with time , and accordingly liquid 1 and liquid 2 can indeed be differentiated by the fraction of locally favoured structures @xmath6 . here we note that liquid 1 is a stable high - temperature liquid and liquid 2 is a low - temperature liquid , which usually exists in a glass state ( glass 2 ) for tpp . this supports the two - order - parameter model of llt @xcite . however , the controversy has still remained due to the lack of direct experimental evidence for the presence of two liquid states and the reversibility of llt . more importantly , no one have succeeded in avoiding nano - crystallization so far , which is due to an intrinsic difficulty associated with the fact that all the llts reported for molecular liquids such as tpp and water take place in a supercooled state below the melting point . to firmly establish the llt scenario , it is desirable to show the reversibility of llt without suffering from any crystallization . for a heating rate less than 1 k s@xmath7 , however , the reverse llt is hidden behind crystallization , even if it exists , since crystallization takes place immediately after the glass 2-to - liquid 2 transition during heating @xcite . terashima et al . observed an endothermic ( heat absorbing ) peak upon heating , which they attributed to the reverse llt on the basis of the heating rate dependence of the onset temperature of the peak @xcite . however , if this is the reverse llt , we should also observe a glass 2-to - liquid 2 transition during heating before glass 2 goes back to liquid 1 , since the process of the reverse llt should occur only in a liquid state and not in a glass state . but they reported only one endothermic peak . furthermore , with a slow heating rate employed in the previous studies , it is impossible to access the entire reverse process from the glacial phase to liquid 1 due to the interference by crystallization . thus , it is still unclear whether the peak is due to the melting of nano - crystals , the glass 2-to - liquid 2 transition , and/or the reverse llt process . in this article , we aim at not only showing the reversibility of llt but also confirming the coexistence of the two ` liquid ' states in the course of the transition in an unambiguous manner . these are crucial for proving that the transition takes place between two distinct liquid states . to this end , we apply ultra high - speed ( flash ) dsc ( differential scanning calorimetry ) , which can provide a heating rate more than 4 orders of magnitude higher than that of conventional dsc ( see methods ) . this allows us to avoid crystallization upon heating and to directly access the reverse process of llt without the interference by crystallization .

the range @xmath133 is commonly regarded as a bound in spectral reconstructions . rather than being a fundamental limit , however , this bound only holds for certain classes of control sequences . here , we show that the ultimate bound is set by the _ time resolution _ of the control sequence , @xmath70 . specifically , @xmath134 , with sequence @xmath135 having duration @xmath136 , where @xmath137 and @xmath138 . for example , all @xmath139 and @xmath140 for the family we construct from different orders of cdd , whereas the @xmath141 are different positive integers and the @xmath142 are rational multiples of @xmath40 for the cpmg sequences used in the protocol of alvarez and suter . each sequence @xmath41 has @xmath143 pulses at times @xmath144 . define @xmath145 and @xmath146 . in practice , the times @xmath147 are constrained by the time resolution @xmath70 : since times less than @xmath70 are not resolvable , all times must be integer multiples of @xmath70 , say @xmath148 for @xmath149 . consequently , the time separations between any two times @xmath150 and @xmath151 must also be integer multiples of @xmath70 . we denote the time separations by @xmath152 for @xmath153 , @xmath154 . for simplicity , consider the task of reconstructing the spectrum of a stationary gaussian noise process . this requires measuring @xmath56^m}$ ] , the decay parameter after @xmath44 repetitions of the control sequence @xmath41 . ( [ eq::chim ] ) in the main text enables us to relate @xmath56^m}$ ] to the psd evaluated at a set of harmonic frequencies @xmath155 for @xmath156 . in the gaussian case , eq . ( [ eq::chim ] ) reduces to @xmath52^m}\approx & -\frac { m}{2t_p}\sum_{h=0}^{h_p } m_{1}\!\!\left(\frac{2\pi h}{t_p}\right ) \big|f_p\!\left(\frac{2\pi h}{t_p}\right)\notag \\ = & -\frac{m}{2t_p}\sum_{h=0}^{h_p } m_{1}\!\!\left(\frac{2\pi n_ph}{t}\right ) \big|f_{p}\!\left(\frac{2\pi n_ph}{t}\right)\big|^2s\!\left(\frac{2\pi n_ph}{t}\right ) \label{eq::chim2}.\end{aligned}\ ] ] inverting the system of linear equations formed by eq . ( [ eq::chim2 ] ) for each @xmath41 in @xmath157 enables one to reconstruct the psd at the base harmonics " , integer multiples of @xmath158 . non - degeneracy of this linear system depends on the periodicity of the ffs . the periodicity of sequence @xmath41 is determined by the oscillatory component of @xmath159 , which is given by @xmath160 - 2(-1)^{n_p}\text{cos}[(t_{n_p+1}-t_0)\omega]\label{eq::ffnumerator}\\ & -4\sum_{j=1}^{n_p}(-1)^{j+n_p}\text{cos}[(t_j - t_{n_p+1})\omega]+4\sum_{j=1}^{n_p}(-1)^j\text{cos}[(t_j - t_0)\omega]\notag . \end{aligned}\ ] ] note that the last two terms in this expression cancel when @xmath41 is a time - symmetric sequence . in an abuse of notation , define @xmath161 , and @xmath162 . for integer @xmath163 , these definitions reduce to the ordinary gcf and lcm , respectively . in the frequency domain , the periodicities of the cosine terms in eq . ( [ eq::ffnumerator ] ) are determined by the time separations @xmath164 . from these cosine terms , the periodicity of @xmath165 is @xmath166 , where @xmath167 here , @xmath168 when @xmath41 is time - symmetric and @xmath169 otherwise . note that the arguments of the cosine terms in eq . ( [ eq::ffnumerator ] ) are even multiples of @xmath10 when @xmath170 . consider now the half periodicity " , @xmath171 . as expected , the oscillatory component is symmetric about @xmath172 , i.e. @xmath173 for all @xmath174 . we now extend this notion of periodicity to a set of control sequences @xmath157 . define the half periodicity of @xmath157 as @xmath175 where @xmath176 note that @xmath177 for all @xmath174 and @xmath135 . because @xmath178 divides all time separations by an integer , @xmath179 for some @xmath180 , implying @xmath181 . in the case that @xmath182 is even , @xmath183 are base harmonics . this implies @xmath184 and @xmath185 ) for each @xmath135 is degenerate and non - invertible when @xmath186 . similarly , if @xmath182 is odd , then @xmath187 are base harmonics and latexmath:[\ ] ] by measuring @xmath56^m}$ ] for a set of control sequences and inverting the system of linear equations formed by eq . ( [ eq::sbchi ] ) , we can obtain the effective spectrum and trispectrum . note that with no prior information about the noise , deciding which form of @xmath56^m}^{(4)}$ ] to use will be an iterative process . for example , from a single set of measurements one can reconstruct both @xmath280 using the entangled form and @xmath290 using the separable form . these reconstructions can then be used to predict the dynamics of the qubit under free decay . by comparing the predictions to actual measurements of the qubit under free decay , one can decide which model best describes the noise , as mentioned in the text .

we introduce open - loop quantum control protocols for characterizing the spectral properties of non - gaussian noise , applicable to both classical and quantum dephasing environments . the basic idea is to engineer a multi - dimensional frequency comb via repetition of suitably designed pulse sequences , through which the desired high - order noise spectra may be related to observable properties of the qubit probe . we prove that access to a high time resolution is key to achieve spectral reconstruction over an extended bandwidth , overcoming limitations of existing schemes . non - gaussian spectroscopy is demonstrated for a classical noise model describing quadratic dephasing at an optimal point , as well as a quantum spin - boson model out of equilibrium . in both cases , we obtain spectral reconstructions that accurately predict the qubit dynamics in the non - gaussian regime . accurately characterizing the spectral properties of environmental noise in open quantum systems has broad practical and fundamental significance . within quantum information processing , this is a prerequisite for optimally tailoring the design of quantum control and error - correcting strategies to the noisy environment that qubits experience , and for testing key assumptions in fault - tolerance threshold derivations @xcite . from a physical standpoint , precise knowledge of the noise is necessary for quantitatively modeling and understanding open - system dynamics , with implications ranging from the classical - to - quantum transition to non - equilibrium quantum statistical mechanics and quantum - limited metrology @xcite . quantum noise spectroscopy seeks to characterize the spectral properties of environmental noise by using a controlled quantum system ( a qubit under multi - pulse control in the simplest case ) as a dynamical probe @xcite . in recent years , interest in quantum noise spectroscopy has heightened thanks to both improved theoretical understanding of open - loop controlled dynamics in terms of transfer filter - function ( ff ) techniques @xcite and experimental validation in different qubit platforms . in particular , quantum control protocols based on dynamical decoupling ( dd ) have been successfully implemented to characterize noise properties during memory and driven evolution in systems as diverse as solid - state nuclear magnetic resonance @xcite , superconducting @xcite and spin @xcite qubits , and nitrogen vacancy centers in diamond @xcite . despite the above advances , existing noise spectroscopy protocols suffer from several disadvantages . notably , they are restricted to _ classical , gaussian _ phase noise . while dephasing ( @xmath0- ) processes are known to provide the dominant decoherence mechanism in a variety of realistic scenarios , the gaussianity assumption is _ a priori _ far less justified . on the one hand , the gaussian approximation tends to break down in situations where the system is strongly coupled with an environment consisting of discrete degrees of freedom such as for @xmath1 noise , as ubiquitously encountered in solid - state devices @xcite . even for environments well described by a continuum of modes , non - gaussian noise statistics may be generally expected away from thermal equilibrium , or whenever symmetry considerations forbid a linear coupling @xcite . in all such cases , accurate noise spectroscopy mandates going beyond the gaussian regime . in this paper , we introduce open - loop control protocols for characterizing _ stationary , non - gaussian dephasing _ using a qubit probe . our approach is applicable to classical noise environments and to a paradigmatic class of open quantum systems described by linearly coupled oscillator environments as long as all relevant noise spectra obey suitable smoothness assumptions . while we build on the noise spectroscopy by sequence repetition proposed by alvarez and suter @xcite , our central insight is to leverage the simple structure of ffs in a purely dephasing setting to establish the emergence of a frequency comb for arbitrary high - order noise spectra ( so - called @xmath2 ) , paving the way to the desired multi - dimensional spectral estimation . we first demonstrate the power of our approach for gaussian noise , where we extend the range of spectral reconstruction over existing protocols . in the non - gaussian regime , we reconstruct the spectra associated with the leading high - order cumulants of the noise , absent in the gaussian limit . quantitative prediction of the qubit free evolution in the presence of these non - gaussian environments reveals clear dynamical signatures in both the classical and quantum case . _ control setting and noise polyspectra. _ we consider a qubit @xmath3 coupled to an uncontrollable environment ( bath ) @xmath4 . in the interaction picture with respect to the bath hamiltonian , @xmath5 , and the qubit hamiltonian @xmath6 , the joint system is described by @xmath7 , where the first term accounts for the bath - induced dephasing and @xmath8 is the external open - loop control , acting non - trivially on the qubit alone . for a classical bath , @xmath9 is a stochastic noise process , whereas @xmath9 is a time - dependent operator for a quantum bath . the applied control consists of repeated sequences of @xmath10-pulses ( say , about @xmath11 ) , which for simplicity we take to be instantaneous . after transforming to the interaction picture associated with @xmath8 , the joint hamiltonian becomes @xmath12 , where the switching function " @xmath13 changes sign between @xmath14 with every @xmath10-pulse applied to the qubit . the effect of dephasing is seen in the dynamics of the qubit s coherence element , which we may express in terms of bath - operator cumulants . specifically , @xmath15 , where the decay parameter and phase angle are respectively given by : @xmath16 where the @xmath17th - order noise cumulant @xmath18 depends on the bath correlation functions @xmath19 , @xmath20 , and @xmath21 denotes an ensemble average for a classical bath or an expectation value with respect to the initial bath state , @xmath22 , in the quantum case . for zero - mean gaussian noise , @xmath23 except for @xmath24 thus , gaussian noise gives _ no phase evolution_. for non - gaussian noise , higher - order even ( odd ) cumulants contribute to decay ( phase evolution ) , respectively . for stationary noise , where @xmath25 is a function of the time separations @xmath26 , @xmath27 , the noise spectral properties are fully characterized by the fourier transforms of the cumulants with respect to @xmath28 . using the compact notation @xmath29 , the _ @xmath30th - order @xmath31 _ is defined as @xmath32 where @xmath33 is the familiar power spectral density ( psd ) , and @xmath34 , @xmath35 are known as the `` bi - spectrum '' and `` tri - spectrum '' , respectively . for all orders , @xmath36 is a smooth @xmath30-dimensional surface when the noise is classical and ergodic @xcite . more generally , @xmath37 may depend on fewer than @xmath30 time separations , leading to the presence of delta functions in @xmath38 . all polyspectra possess a high degree of symmetry , _ irrespective of the noise_. that is , @xmath38 is fully specified in all frequency space by its value on a particular subspace , @xmath39 , known as the _ principal domain _ @xcite . _ noise spectroscopy protocol. _ our objective is to characterize not only the psd but the polyspectra . we accomplish this by adapting the dd noise spectroscopy protocol proposed in @xcite for gaussian noise . this protocol relies on repetitions of identical _ base sequences _ , whose duration ( `` cycle time '' ) we shall denote by @xmath40 . following @xcite , the effect of a base control sequence @xmath41 in the frequency domain is characterized by a single _ fundamental ff _ , @xmath42 . if @xmath43 , direct calculation shows that @xmath44 repetitions of @xmath41 yield @xmath45^m}^{(k ) } \!=&\!\!\!\int_{\mathbb{r}^{k-\!1}}\!\!\!\!\!\!\!d\vec{\omega}_{k-\!1}\!\ ! \prod_{j=1}^{k-\!1}\!\!f_p(\omega_j)\frac{\text{sin}(m\omega_jt/2)}{\text{sin}(\omega_jt/2)}\notag \\&\times f_p(-|\vec{\omega}_{k-\!1}|)\frac{\text{sin}(m|\vec{\omega}_{k-\!1}|t/2)}{\text{sin}(|\vec{\omega}_{k-\!1}|t/2)}\frac{s_{k-\!1}(\vec{\omega}_{k-\!1})}{(2\pi)^{k-\!1 } } , \label{eq::ftcumulantm}\end{aligned}\ ] ] the key to extending the protocol in @xcite beyond gaussian noise ( @xmath46 ) is to realize that _ repetition produces a multi - dimensional frequency comb for all orders _ , namely , @xmath47\frac{\text{sin}(m|\vec{\omega}_{k-1}|t/2)}{\text{sin}(|\vec{\omega}_{k-1}|t/2 ) } \label{eq::hypercomb } \\&\approx\!m\prod_{j=1}^{k-1}\!\!\big[\frac{2\pi}{t}\!\!\sum_{n_j\!=-\!\infty}^{\infty}\!\!\!\delta\big(\omega_{j}\!-\!\frac{2\pi n_j}{t}\big)\big]\notag , \quad m \gg1 , \forall k , \end{aligned}\ ] ] provided that @xmath48 in eq . ( [ eq::ftcumulantm ] ) is a smooth function . thanks to the `` hyper - comb '' in eq . ( [ eq::hypercomb ] ) , obtaining the polyspectra becomes an inverse problem . substituting eq . ( [ eq::hypercomb ] ) into eq . ( [ eq::ftcumulantm ] ) produces a linear equation that couples the polyspectra and the ffs evaluated at the harmonic frequencies @xmath49 , @xmath45^m}^{(k)}\!\!=\!\!\!\!\!\!\!\!\!\sum_{\vec{h}_{k-\!1}\in { \cal h}_{k-\!1}}\!\!\!\!\!\!\!\!\frac{m}{t^{k-\!1}}\!\!\prod_{j=1}^{k-1}\!\!f_p(h_j ) f_p(-|\vec{h}_{k-\ ! 1}|)s_{k-\!1}(\vec{h}_{k-\!1 } ) \label{eq::ftlinear}.\end{aligned}\ ] ] to obtain a finite linear equation , we need to truncate the above sum to a finite set @xmath50 . with no prior knowledge of the noise , it suffices to consider @xmath51 in the principal domain of the polyspectrum . truncating the expression in eq . ( [ eq::ftlinear ] ) enables us to relate the sampled polyspectra to experimentally observable dynamical quantities : @xmath52^m } \approx&\sum_{\ell=1}^\infty\frac{(-1)^\ell m}{(2\ell)!\,t^{2\ell-\!1}}\hspace*{-3 mm } \sum_{\vec{h}_{2\ell-\!1}\in\omega_{2\ell-\!1}}\!\!\!\!\!m_{2\ell-\!1}(\vec{h}_{2\ell-\!1 } ) \nonumber \\ & \times\prod_{j=1}^{2\ell-\!1}\!\!f_p(h_j ) f_p(-|\vec{h}_{2\ell-\!1}|)s_{2\ell-1}(\vec{h}_{2\ell-\!1 } ) , \label{eq::chim } \\ \phi_{[p]^m } \approx&\sum_{\ell=1}^\infty\frac{(-1)^\ell m}{(2\ell+1)!\,t^{2\ell}}\sum_{\vec{h}_{2\ell}\in\omega_{2\ell}}m_{2\ell}(\vec{h}_{2\ell } ) \nonumber \\ & \times\prod_{j=1}^{2\ell}f_p(h_j ) f_p(-|\vec{h}_{2\ell}|)s_{2\ell}(\vec{h}_{2\ell } ) , \label{eq::phim}\end{aligned}\ ] ] where the multiplicity @xmath53 accounts for the symmetry of the polyspectrum . whenever the contributions from high - order multi - point correlation functions are negligible ( e.g. , for sufficiently small evolution time and/or noise strength ) , the cumulant expansion in eqs . ( [ eq::chim])-([eq::phim ] ) may be truncated at a finite @xmath54 . if @xmath55 terms remain after truncation , experimentally measuring @xmath56^m}$ ] ( @xmath57^m}$ ] ) for at least @xmath55 control sequences creates a system of linear equations , that can be inverted to obtain the odd ( even ) polyspectra up to order @xmath58 ( @xmath59 ) @xcite . _ base sequence construction. _ in the original noise spectroscopy protocol of @xcite , a _ fixed _ base sequence is used ( cpmg , after carr , purcell , meiboom , and gill ) , with cycle times varying from @xmath40 to @xmath60 , where @xmath61 and @xmath62 is the minimum time between pulses . while this produces a well - conditioned linear inversion , both the number of distinct control sequences and the range of spectral reconstruction are limited in particular , @xmath63 for a minimum allowed @xmath64 . the use of a fixed dd sequence has an additional disadvantage : cpmg refocuses static noise ( @xmath65 , hence the `` filtering order '' is non - zero @xcite ) , precluding reconstruction at any point in frequency space containing a zero , a substantial information loss for higher - dimensional polyspectra . non - gaussian noise spectroscopy demands a _ large number of sequences with spectrally distinct ffs _ , including some with zero filtering order . we generate a family of base sequences satisfying these requirements by using different orders of concatenated dd , cdd@xmath66 : namely , not only cpmg ( @xmath67 ) , but also durations of free evolution ( @xmath68 ) , up to maximum dd order @xmath69 . the presence of free evolution permits sequences with zero filtering order , enabling the polyspectra to be reconstructed at points containing a zero . specifically , let a _ fixed _ cycle time @xmath40 be expressed in terms of a minimum _ time resolution _ @xmath70 , @xmath71 , where @xmath72 . while all pulse times will be integer multiples of @xmath70 , @xmath70 and @xmath62 are two independent constraints a priori , with @xmath73 in typical settings . if @xmath74 is an integer partition of @xmath75 , we place a cdd@xmath66 sequence into the @xmath76th interval , of duration @xmath77 , subject to the condition that no two pulses are separated by less than @xmath62 . as shown in the supplement @xcite , the range of spectral reconstruction is bounded by @xmath78 . a high resolution ( small @xmath70 ) is key to generate sequences with _ incommensurate _ periodicities , making it possible to achieve spectral reconstruction over an extended range . repetitions of control sequences with @xmath79s and @xmath80 . for our protocol , we employ 25 base sequences assembled from cdd@xmath66 , @xmath81 . the spectrum is a sum of lorentzians peaks , @xmath82 ^ 2 } + { w_2}/\{1 + [ 8\,(\text{sign}(\omega)\omega - d)/\omega_c)^2]\}$ ] , where @xmath83 , @xmath84 , @xmath85 , and @xmath86 controls the offset of the high - frequency peaks , @xmath87 ( a ) , @xmath88 ( b ) , @xmath89 ( c ) . as the original protocol can only reconstruct @xmath90 up to @xmath91 ( dashed vertical lines ) , it can not see " the high - frequency peaks in ( b)-(c ) , which results in instability at lower frequencies . ] the added capabilities of our control sequences may be appreciated already for spectroscopy of classical gaussian noise , see fig . [ fig::lorentzian3pitau ] . in this case , eq . ( [ eq::chim ] ) truncates exactly at @xmath92 ; this produces a system of linear equations relating the desired psd to @xmath56^m}$ ] , which we obtain numerically for each control sequence . in addition to accurately reconstructing the larger peaks over the expanded range @xmath93 , our protocol successfully resolves the small peak at @xmath94 , thanks to inclusion of control sequences with zero filtering order . with different degrees of gaussianity [ see text ] , and same spectrum @xmath95 for @xmath96 as in fig . [ fig : classicalbispectrum ] . curves are ordered according to decreasing degree of non - gaussianity : @xmath97 ( blue solid ) , @xmath98 ( red dashes ) and @xmath99 ( green dots ) . for the fully non - gaussian @xmath97 case , we have used the reconstructed spectrum and bi - spectrum [ fig . [ fig : classicalbispectrum ] ] to predict the qubit decay and phase evolution ( blue asterisks ) , showing excellent agreement with the theoretical evolution , computed up to the fifth - order noise cumulant . ] _ non - gaussian spectral reconstructions. _ we now return to our main goal , namely characterizing non - gaussian polyspectra . as a first example , we consider a classical square noise " process arising from a quadratic coupling to a gaussian source , as encountered in superconducting qubits operating at an optimal working point @xcite . that is , @xmath100 , where @xmath96 is a zero - mean gaussian process , and @xmath101 $ ] interpolates between gaussian ( @xmath102 ) and fully non - gaussian ( @xmath97 ) regimes . truncating eq . ( [ eq::phim ] ) at the leading @xmath92 term allows us to reconstruct the bi - spectrum @xmath34 from numerically determined values of @xmath57^m}$ ] . here , the relevant principal domain @xmath103 is an octant bounded by @xmath104 and @xmath105 . reconstructing 35 points in @xmath106 enables us to obtain @xmath34 at 325 points in @xmath107 . representative results for the actual vs. reconstructed bi - spectrum at @xmath97 are shown in fig . [ fig : classicalbispectrum](a)-(b ) . the relative error in [ fig : classicalbispectrum](c ) indicates very good agreement at the interior points , but larger error in the tails . because there is minimal spectral concentration in the tails , however , this error has little effect on the qubit dynamics . as fig . [ fig::fid ] shows , excellent agreement is found between the theoretical phase evolution and the one predicted by the reconstructed bi - spectrum . extending quantum noise spectroscopy methods to quantum environments entails qualitatively new challenges because non - gaussian statistics ensues now from the _ combined effect of the bath operators @xmath9 and the initial bath state _ @xmath22 , and no general characterization of quantum polyspectra ( and their smoothness properties ) is available to the best of our knowledge . we take a first step in this direction by focusing on linearly coupled spin - boson environments , in which case @xmath108 and @xmath109 , where @xmath110 are canonical bosonic operators and both @xmath111 have units of ( angular ) frequency . for a general quantum bath , the noise is stationary if and only if @xmath112=0 $ ] . this prevents non - zero odd cumulants in the spin - boson model , implying that the qubit undergoes no phase evolution . given any stationary , non - gaussian bath state @xmath22 we can , using our protocol , reconstruct spectral quantities associated with the first two leading - order even cumulants , @xmath90 and @xmath35 . although @xmath90 is _ asymmetric _ about @xmath94 , the fact that arbitrary ffs enter through even combinations implies that we can only reconstruct its symmetric component , or effective spectrum " , @xmath113/2 $ ] . @xmath114 is the quantity relevant to the qubit dynamics . as shown in @xcite , the tri - spectrum for any _ non - separable _ , stationary initial bath state has the form @xmath115.$ ] because the hypercomb approximation holds only if @xmath35 is smooth , we can not directly reconstruct it . we can , however , reconstruct the effective tri - spectrum " @xmath116 , provided it is smooth . due to the delta functions in @xmath35 , the terms in eq . ( [ eq::chim ] ) associated with the tri - spectrum differ by a constant factor in the spin - boson case . the modified equations are derived in @xcite , along with similar equations for _ separable _ stationary initial states . in the absence of prior information about @xmath22 , comparison between predictions based on the two resulting reconstructions will enable the correct effective tri - spectrum to be inferred . for illustration , we choose here @xmath22 to be a non - gaussian , non - separable state corresponding to far - from - equilibrium conditions , and simultaneously reconstruct @xmath117 and @xmath118 by numerically determining @xmath56^m}$ ] and inverting the appropriate system of linear equations @xcite . to test the accuracy of our reconstructions , we again predict the dynamics of the qubit under free evolution . as shown in fig . [ fig::sb](a ) , taking into account the non - gaussianity of the noise by reconstructing both the effective spectrum and tri - spectrum improves the prediction by almost an order of magnitude in time . because the non - gaussian prediction only uses spectral quantities associated with the second and fourth cumulants , however , it fails when the fourth cumulant becomes comparable in size to the second , indicating that the higher - order cumulants can no longer be neglected ( see also fig . [ fig::sb](b ) ) . ( b ) , reconstructed effective tri - spectrum ( c ) . the non - gaussian initial bath state is @xmath119 , where @xmath120 , @xmath121 are thermal states at temperatures @xmath122k , @xmath123k . ohmic spectral density @xmath124 is assumed , with @xmath125 nhz , @xmath126 khz . the curves in ( a ) represent theoretical decay ( black solid ) , decay predicted by reconstructing @xmath117 and @xmath118 ( grey asterisks ) , and decay predicted by approximating the noise as gaussian and reconstructing @xmath117 only ( teal squares ) . the non - gaussian prediction in ( b ) fails when @xmath127 and @xmath128 become comparable . all reconstructions used @xmath129 repetitions of 21 base sequences composed of cdd@xmath130 , with @xmath131s , @xmath132 . ] _ conclusion. _ we introduced control protocols for characterizing the high - order spectra associated with non - gaussian dephasing on a qubit probe coupled to a classical or a quantum bosonic environment . our approach overcomes limitations of existing protocols , allowing in particular for spectral reconstruction over an extended bandwidth , which is of independent interest for quantum sensing applications . our work also points to the need for a deeper understanding of high - order quantum noise spectra beginning from more complex dephasing settings described by non - linear spin - boson models or spin baths . we expect implementation of our protocols to be within reach for various device technologies , in particular transmon or flux qubits @xcite , where they may help shed light onto the microscopic origin of the noise itself . we thank m. biercuk , w. oliver , s. gustavsson , and f. yan for valuable discussions . work at dartmouth was supported from the us aro under contract no . w911nf-14 - 1 - 0682 and the constance and walter burke special projects fund in quantum information science . gaps acknowledges support from the arc centre of excellence grant no . ce110001027 . 16ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop _ _ ( , , ) @noop @noop @noop * * , ( ) @noop * * , ( ) @noop @noop @noop @noop * * , ( ) @noop in @noop _ _ ( , ) p. @noop * * , ( ) @noop @noop @noop * * , ( )

determining the absolute values of the three neutrino masses is a subject of great interest . historically , pauli introduced the concept of the neutrino in order to explain why the energy of the emitted electron in beta decay did not have a single value as expected for a two body final state , but had a spectrum of energies ( for different observations of the process ) . fermi then suggested experimentally determining the mass of the neutrino by precisely measuring the electron s energy spectrum in the endpoint region . this procedure has yielded steadily decreasing upper bounds on neutrino mass for more than half a century . it has however turned out to be difficult to reach the range suggested by the cosmology bound @xcite for the three different neutrinos now known to exist : @xmath0 this bound is based on the gravitational force exerted by neutrinos in the universe rather than directly from neutrino kinematics .. however , heroic efforts @xcite - @xcite are continuing to make steady progress on finding the neutrino masses using the kinematics of the tritium beta decay reaction . one of course has the associated problem of finding the masses of all three neutrino mass eigenstates . the neutrino oscillation experiments @xcite - @xcite have determined these mass differences up to an ambiguity of either a normal " or an inverse " hierarchy . specifically , @xmath1 where the numbers have been taken from the recent review @xcite . of course , the two hierarchies correspond to the two possible signs of @xmath2 . evidently , measuring any combination of three neutrino masses which is independent of @xmath3 and @xmath4 should determine , for each hierarchy choice , the values of all three neutrino masses . here we would like to discuss some aspects related to the possibility of finding the separate ( or absolute " ) neutrino masses with the help of experiments measuring neutrino velocities by timing laboratory - made neutrino beams traveling from one point to another . at the moment , and for the near future , these are thought experiments " . however , thinking about them , raises a number of interesting questions . if there were just a single neutrino it is clear that we could trivially " find its mass , @xmath5 by measuring its velocity , @xmath6 as well as its energy , @xmath7 and using einstein s formula : @xmath8 where we are working in units where the velocity of light , @xmath9 and @xmath10 , appropriate to neutrinos traveling just slightly less than the speed of light . for definiteness , in the realistic 3-neutrino world , we assume a neutrino beam initiated in association with muons and detected , after traveling a measured time and distance , by produced muons . we designate this as a mu - neutrino type beam . now we must take account of the fact that the laboratory neutrino beam in this setup corresponds to working with a neutrino flavor " eigenstate ( of muonic type ) rather than a mass eigenstate . a prescription is therefore required to define what might be called the averaged neutrino mass " @xmath5 which enters into eq . ( [ einfor ] ) when we want to use it in the 3-flavor world . in quantum mechanics the mu - type neutrino corresponds to a linear combination of the three mass eigenstates : @xmath11 where the @xmath12 s are approximately the elements of the leptonic mixing matrix . in models , say for the case of majorana neutrinos , the mixing results from bringing anoriginal " mass matrix @xmath13 to diagonal form @xmath14 , @xmath15 where @xmath16 is unitary . the to - be - measured " mass , m will thus be interpreted @xcite as the averaged " quantity , @xmath17 where the repeated index , @xmath18 should be summed over . we also have assumed for simplicity that the @xmath19 s are real . as a check of this formula , one notes that if the neutrinos did nt mix we would have @xmath20 and @xmath21 . present experimental information on the angles " which parameterize the matrix @xmath16 are summarized in ref . @xcite translating these angles to the needed matrix elements , @xmath22 yields the numerical values with estimated errors : @xmath23 eq . ( [ avmass ] ) then explicitly gives the to - be - measured quantity , @xmath5 as a different combination of the masses than those in eq . ( [ massdif ] ) . thus combining the three equations we can find the absolute neutrino masses for each hierarchy choice . for example , we may eliminate @xmath24 and @xmath25 in terms of @xmath26 by using eqs . ( [ massdif ] ) to get @xmath27 here the plus sign corresponds to the normal hierarchy choice and the minus sign corresponds to the inverted hierarchy choice . for the normal hierarchy we can , using this equation , obtain the curve in fig 1 while for the inverted hierarchy we obtain the curve in fig . 2 . thus if an average mass , m is measured by a long baseline time of flight experiment , we can read @xmath26 from this curve and find also ( from the known differences above ) , both @xmath24 and @xmath25 . this should be done for both hierarchy assumptions ; the present method does not determine which hierarchy is the one nature chooses . of course , there are some important experimental errors introduced in this determination . it is expected that further experimental measurements of neutrino oscillations will substantially improve the experimental accuracy . this procedure thus appears to be able to reasonably solve the problem of relating the averaged " mass , @xmath5 measured by using eq.([avmass ] ) to the three neutrino mass eigenvalues @xmath28 . however , we are not done since we should pay some attention to the accuracy to which @xmath5 itself can be determined . it should be remarked that the need to take mixtures of mass eigenstates into account is also evident in the electron endpoint of beta decay method for absolute neutrino mass determination . in that case the @xmath29 s appear instead of the @xmath22 s ; see for example eq . ( 21 ) of @xcite . next , we discuss the needed determination of the averaged mass , @xmath5 by a long baseline time of flight measurement . first the mu - neutrino , @xmath30 must be produced . we would like to select the @xmath30 to have as small an energy as possible in order that it travel at as low a speed , @xmath6 as possible . an at least conceptually convenient source in this context might use the decay process : @xmath31 by selecting the @xmath32 and @xmath33 momenta one can reduce the energy of the @xmath30 and hence its speed , @xmath6 . that would make a precise time measurement easier . from eq . ( [ einfor ] ) , the deviation of the neutrino s speed from that of light is given by : @xmath34 if one were to select the neutrino energy to be 1 mev , this would result in @xmath35 to be of the order of @xmath36 . that corresponds to a deviation from the velocity of light in the fourteenth decimal place , which does not seem practical . however , if one were able to get enough data with a neutrino energy selected to be 1 kev , @xmath35 would be of the order of @xmath37 which should be measurable . we have shown how measurements of , for example , a mu - neutrino beam velocity and energy can be used to help in the determination of the absolute masses of the presently known three neutrinos . this could be a supplement to the determinations obtained using the endpoint of tritium beta decay method . of course , an important question is the accuracy which can be achieved . in addition to the accuracy of the velocity and energy measurements of the beam it is important to increase the accuracy of the neutrino mass differences and the lepton mixing matrix elements obtained via neutrino oscillation experiments . a recent review of these is given in @xcite . -.5 cm this work was supported in part by the u. s. doe under contract no . de - fg-02 - 85er 40231 . one of us ( js ) would like to thank prof . f. sannino and the members of the cp@xmath38 origins group at southern denmark university for their helpful discussions , warm hospitality and partial support ..

we discuss the possibility of using experiments timing the propagation of neutrino beams over large distances to help determine the absolute masses of the three neutrinos .

in @xcite wachs constructs a basis for the homology of the partition lattice @xmath10 via a certain natural `` splitting '' procedure for permutations . this basis has very favorable properties with respect to the representation of the symmetric group @xmath11 on @xmath12 , a representation that had earlier been studied by stanley @xcite , hanlon @xcite and many others . it also is the shelling basis for a certain el - shelling of the partition lattice given in ( * ? ? ? * section 6 ) . this basis has connections to the free lie algebra as well ; see @xcite . we now give a brief description of the splitting basis of @xcite . for each @xmath13 , let @xmath14 be the subposet of @xmath10 consisting of partitions obtained by splitting @xmath15 . in figure 1 the subposet @xmath16 of @xmath17 is shown . each poset @xmath14 is isomorphic to the face lattice of an @xmath18-dimensional simplex . therefore @xmath19 , the order complex of the proper part of @xmath14 , is an @xmath20-sphere embedded in @xmath21 , and hence it determines a fundamental cycle @xmath22 . in @xcite it is shown that a certain subset of @xmath23 forms a basis for @xmath24 ; namely , the set of all @xmath25 such that @xmath15 fixes @xmath26 . the partition lattice is the intersection lattice of the type @xmath0 coxeter arrangement . the original motivation for this paper was to explain and generalize to other coxeter groups , the splitting basis for @xmath10 . taking a geometric point of view we give such an explanation , which then leads to the construction of `` splitting bases '' also for the intersection lattices of coxeter arrangements of types @xmath1 and @xmath2 and of some interpolating arrangements . our technique is general in that it gives a way to construct a basis for the homology of the intersection lattice of any real hyperplane arrangement . the intersection lattice of the type @xmath1 coxeter arrangement is isomorphic to the signed partition lattice @xmath27 . its elements are signed partitions of @xmath28 ; that is partitions of @xmath28 in which any element but the smallest element of each nonzero block can be barred . for each element @xmath15 of the hyperoctahedral group @xmath29 , we form a subposet @xmath14 of @xmath27 consisting of all signed partitions obtained by splitting the signed permutation @xmath15 . in figure 2 the subposet @xmath30 of @xmath31 just as for type @xmath0 , it is clear that each subposet @xmath14 determines a fundamental cycle @xmath25 in @xmath32 . it is not clear , however , that the elements @xmath25 , @xmath33 , generate @xmath34 ; nor is it clear how one would select cycles @xmath25 that form a basis for @xmath34 . our geometric technique enables us to identify a basis whose elements are those @xmath25 for which the right - to - left maxima of @xmath15 are unbarred . we will now give a somewhat more detailed description of the content of the paper . the proper setting for our discussion is that of real hyperplane arrangements , or ( even more generally ) oriented matroids . let @xmath3 be an arrangement of linear hyperplanes in @xmath4 . we assume that @xmath3 is essential , meaning that @xmath35 . the intersection lattice @xmath36 is the family of intersections of subarrangements @xmath37 , ordered by reverse inclusion . it is a geometric lattice , so it is known from a theorem of folkman @xcite that @xmath38 and @xmath39 for all @xmath40 , where @xmath41 . in fact , the order complex @xmath42 has the homotopy type of a wedge of @xmath43-spheres . there are many copies of the boolean lattice @xmath44}$ ] ( or equivalently , the face lattice of the @xmath45-simplex ) embedded in every geometric lattice of length @xmath46 . each such boolean subposet determines a fundamental cycle in homology . in @xcite bjrner gives a combinatorial method for constructing homology bases using such boolean cycles . this method , which in its simplest version is based on the so called `` broken circuit '' construction from matroid theory , is applicable to _ all _ geometric lattices ( not only to intersection lattices of hyperplane arrangements ) . although the cycles in the splitting basis are boolean , the basis does not arise from the broken circuit construction . it turns out that the splitting basis does arise from the geometric construction in this paper . there is a natural way to associate polytopal cycles in the intersection lattice @xmath47 with regions of the arrangement @xmath48 . these cycles are not necessarily boolean . they are fundamental cycles determined by face lattices of convex @xmath45-polytopes embedded in @xmath47 . we show that these cycles generate the homology of @xmath49 . moreover , we present a way of identifying those regions whose corresponding cycles form a basis . here is a short and non - technical statement of the method . let @xmath50 be an affine hyperplane in @xmath4 which is generic with respect to @xmath3 . the induced affine arrangement @xmath51 in @xmath52 will have certain regions that are bounded . each bounded region @xmath53 is a convex @xmath45-polytope in @xmath50 and it is easy to see that a copy of its face lattice sits embedded in @xmath36 . briefly , every face @xmath54 of @xmath53 is the intersection of the maximal faces containing it , and so @xmath54 can be mapped to the intersection of the linear spans ( in @xmath4 ) of these maximal faces , which is an element of @xmath36 . thus , we have a cycle @xmath55 for each bounded region @xmath53 . a main result ( theorem [ t2 ] ) is that these cycles @xmath56 , indexed by the bounded regions of @xmath57 , form a basis for @xmath58 . the regions of a coxeter arrangement are simplicial cones that correspond bijectively to the elements of the coxeter group . when the geometric method is applied to the intersection lattice of any coxeter arrangement , the cycles in the resulting basis are boolean and are indexed by the elements of the coxeter group that correspond to the bounded regions of a generic affine slice . for type a , when the generic affine hyperplane @xmath50 is chosen appropriately one gets the splitting basis consisting of cycles @xmath25 indexed by the permutations @xmath15 that fix @xmath26 . in figure 3 the intersection of the coxeter arrangement @xmath59 with @xmath50 the bounded regions are labeled by their corresponding permutation . for type b , when the generic affine hyperplane @xmath50 is chosen appropriately , one gets the type @xmath1 splitting basis consisting of cycles @xmath25 indexed by signed permutations @xmath15 whose right - to - left maxima are unbarred . the hyperplane arrangement @xmath60 intersected with a cube is shown in figure 4 . the regions that have bounded intersection with @xmath50 are the ones that are labeled . the labels are the signed permutations whose right - to - left maxima are unbarred . all arguments in the paper are combinatorial in nature , which means that they can be carried out for oriented matroids . so the construction of bases is applicable to geometric lattices of orientable matroids . geometrically this means that we can allow some topological deformation of the hyperplane arrangements . major parts of this work ( sections [ affine ] , [ cent ] and [ braid ] ) were carried out at the hebrew university in 1993 during the jerusalem combinatorics conference . the rest was added in 1998 . it has been brought to our attention that some of the material in sections [ affine ] and [ cent ] shows similarities with work of others ( see e.g. proposition 5.6 of damon @xcite and parts of ziegler @xcite , @xcite ) ; however , there is no substantial overlap or direct duplication . the concept of a _ shellable complex _ and a _ shellable poset _ will be considered known . see @xcite for the definition and basic properties . in particular , we will make use of the _ shelling basis _ for homology and cohomology ( * ? ? ? * section 4 ) . a facet @xmath54 will be called a _ full restriction facet _ with respect to a shelling if @xmath61 , where @xmath62 is the restriction operator induced by the shelling . ( _ remark : _ such facets were called _ homology facets _ in ( * ? ? ? * section 4 ) . ) our notation for posets is that of ( * ? ? ? * section 5 ) . for instance , if @xmath63 is a bounded poset with top element @xmath64 and bottom element @xmath65 then @xmath66 denotes the _ proper part _ of @xmath63 , which is defined to be @xmath67 ; and if @xmath63 is an arbitrary poset then @xmath68 . also , define @xmath69 and @xmath70 . the following simple lemma is a useful devise for identifying bases for homology of simplicial complexes . it is used implicitly in ( * ? ? ? * proof of theorem 2.2 ) and variations of it are used in @xcite . for any element @xmath71 of the chain complex of a simplicial complex @xmath72 and face @xmath54 of @xmath72 , we denote the coefficient of @xmath54 in @xmath71 by @xmath73 . [ l1 ] let @xmath72 be a @xmath46-dimensional simplicial complex for which @xmath74 has rank @xmath75 . if @xmath76 are @xmath46-cycles and @xmath77 are facets such that the matrix @xmath78}$ ] is invertible over @xmath79 , then @xmath76 is a basis for @xmath74 . let @xmath80 . then @xmath81 } = ( 0,\dots,0).\ ] ] since @xmath82}$ ] is invertible , @xmath83 for all @xmath84 . hence @xmath76 are independent over @xmath85 as well as @xmath79 . it follows that @xmath76 forms a basis over @xmath85 . to see that @xmath76 spans @xmath74 , let @xmath71 be a @xmath46-cycle . then @xmath86 where @xmath87 . we have @xmath88 } = ( \langle\rho , f_1\rangle,\dots,\langle\rho , f_t\rangle)\ ] ] it follows that @xmath89}^{-1 } \in \z^t.\ ] ] hence @xmath71 is in the @xmath79-span of @xmath76 . suppose that @xmath90 is a shelling order of the maximal chains of a pure shellable poset @xmath63 of length @xmath91 . let @xmath92 be the set of maximal elements of @xmath63 . recall the following two facts : 1 . for each @xmath93 , a shelling order @xmath94 is induced on the maximal chains of @xmath95 by restricting @xmath90 to the chains containing @xmath96 ( * ? ? ? * prop 4.2 ) . 2 . a shelling order @xmath97 is induced on the maximal chains of @xmath98 as follows . map each maximal chain @xmath99 in @xmath100 to its @xmath90-earliest extension @xmath101 , @xmath102 . note that @xmath103 is injective . now say that @xmath99 precedes @xmath104 in @xmath97 if and only if @xmath105 precedes @xmath106 ( * ? ? ? let @xmath107 and @xmath108 denote the sets of full restriction facets induced by @xmath94 and @xmath97 . recall from ( * ? ? ? * section 4 ) that the shelling @xmath109 induces a basis @xmath110 of @xmath111 which is characterized by the property that @xmath112 for all @xmath113 . [ l2 ] let @xmath63 be a pure poset of length @xmath91 and @xmath92 the set of its maximal elements . suppose that @xmath63 is shellable and acyclic . then 1 . @xmath114 , 2 . @xmath115 is a basis for @xmath116 . we claim that [ claim ] c ( p_<m ) ( c ) = c\{m } c ( p m).let @xmath117 . this means that @xmath118 is contained in an @xmath94-earlier maximal chain of @xmath95 , for every @xmath119 . if @xmath120 with @xmath121 then it would follow that @xmath122 is a full restriction facet of @xmath63 , contradicting the assumption that @xmath63 is acyclic . hence @xmath123 . we can also conclude that @xmath124 . it follows from ( [ claim ] ) that the sets @xmath125 , are disjoint and that @xmath126 the reverse inclusion will be a consequence of the following computations using the mbius function @xmath127 of @xmath128 . since @xmath63 is acyclic we have that @xmath129 hence , @xmath130 for the homology basis of @xmath131 we will use lemma [ l1 ] . order @xmath108 by @xmath97 , and for each @xmath132 , let @xmath133 be defined by @xmath134 . by ( [ claim ] ) , @xmath135 . let @xmath136 be the element of @xmath137 corresponding to @xmath99 . so , @xmath136 is the @xmath138-cycle in @xmath139 with coefficient @xmath140 at @xmath99 and coefficient 0 at all @xmath141 . suppose that @xmath136 has nonzero coefficient at some chain @xmath142 . since @xmath104 must come before @xmath99 in @xmath94 ( the cycle @xmath136 has support on a subset of the chains in @xmath95 that were present at the stage during the shelling @xmath94 when @xmath99 was introduced ) , it follows that @xmath106 precedes @xmath105 in @xmath90 , and hence that @xmath104 precedes @xmath99 in @xmath97 . hence the matrix @xmath143 is lower triangular with @xmath144 s on the diagonal . it now follows from lemma [ l1 ] that @xmath145 is a basis for @xmath146 @xmath147 . let @xmath148 be an arrangement of affine ( or linear ) hyperplanes in @xmath4 . each hyperplane @xmath149 divides @xmath4 into three components : @xmath149 itself and the two connected components of @xmath150 . for @xmath151 , say that @xmath152 if @xmath153 and @xmath154 are in the same component with respect to @xmath149 , for all @xmath155 . this equivalence relation partitions @xmath4 into open cells . let @xmath156 denote the poset of cells ( equivalence classes under @xmath157 ) , ordered by inclusion of their closures . @xmath156 is called the _ face poset _ of @xmath3 . it is a finite pure poset with at most @xmath158 rank levels corresponding to the dimensions of the cells . the maximal elements of @xmath156 are the regions of @xmath159 . see ziegler @xcite for a detailed discussion of these facts . assume in what follows that the face poset @xmath156 has length @xmath46 . we will make use of the following technical properties of the order complex of @xmath156 . [ p1 ] 1 . @xmath156 is shellable . @xmath156 is homeomorphic to the @xmath46-ball . 3 . let @xmath53 be a region of @xmath159 . then @xmath160 if @xmath53 is a bounded region then its closure @xmath161 is a convex @xmath46-polytope , and the open interval @xmath162 is the proper part of the face lattice of @xmath161 . the order complex of @xmath162 , being a simplicial @xmath45-sphere , supports a unique ( up to sign ) _ fundamental @xmath45-cycle _ @xmath163 . let @xmath164 . equivalently , @xmath165 is the poset @xmath156 with its maximal elements ( the regions ) removed . also , let @xmath166 . [ p2 ] 1 . @xmath165 has the homotopy type of a wedge of @xmath45-spheres . 2 . @xmath167 is a basis for @xmath168 . part ( i ) follows from the fact that shellability is preserved by rank - selection ( * ? ? ? 4.1 ) , and that a shellable pure @xmath45-complex has the stated homotopy type . since @xmath169 is ( due to uniqueness ) the shelling basis for @xmath170 when @xmath171 , and @xmath172 when @xmath173 , part ( ii ) follows from lemma [ l2 ] . let @xmath36 denote the _ intersection semilattice _ of @xmath3 . its elements are the nonempty intersections @xmath174 of subfamilies @xmath37 , and the order relation is reverse inclusion . @xmath36 is a pure poset of length @xmath46 . its unique minimal element is @xmath4 ( corresponding to @xmath175 ) , which ( according to convention ) will be denoted by @xmath65 . the minimal elements of @xmath176 are the hyperplanes @xmath177 , and the maximal elements are the single points of @xmath4 obtainable as intersections of subfamilies @xmath37 . @xmath36 is a _ geometric semilattice _ in the sense of @xcite . for each cell @xmath178 , let @xmath179 be the affine span of @xmath180 . the subspace @xmath179 can also be described as follows . by definition , @xmath180 is the intersection of certain hyperplanes in @xmath3 ( call the set of these hyperplanes @xmath181 ) and certain halfspaces determined by other hyperplanes in @xmath3 . then , @xmath182 . this shows that @xmath183 and that @xmath184 . the map @xmath185 is clearly order - reversing , and it restricts to an order - reversing map @xmath186 in various versions , the following result appears in several places in the literature ; see the discussion following lemma 3.2 of @xcite . [ p3 ] the map @xmath187 induces homotopy equivalence of order complexes . we will use the quillen fiber lemma @xcite . this reduces the question to checking that every fiber @xmath188 is contractible , @xmath189 . but by proposition [ p1 ] ( ii ) such a fiber is homeomorphic to a @xmath190-ball , so we are done . the simplicial map @xmath191 induces a homomorphism @xmath192 which ( as a consequence of proposition [ p3 ] ) is an isomorphism . the following is an immediate consequence of propositions [ p2 ] and [ p3 ] . [ t1 ] @xmath193 is a basis of @xmath194 . recall that @xmath163 is the fundamental cycle of the proper part of the face lattice of the convex polytope @xmath161 , for each bounded region @xmath53 . since the map @xmath191 is injective on each lower interval @xmath195 it follows that the cycles @xmath196 are also `` polytopal '' , arising from copies of the proper part of the dual face lattice of @xmath161 embedded in @xmath36 . it is a consequence of theorem [ t1 ] that @xmath197 this enumerative corollary is equivalent to the following result of zaslavsky @xcite : @xmath198 indeed , we have that @xmath199 since @xmath200 is the intersection lattice of a central arrangement and is hence a geometric lattice . since @xmath201 the results are equivalent . our work in this section has the purpose to provide a short but exact route to the results of the following section , in particular to theorem [ t2 ] . in the process , a natural method for constructing bases for geometric semilattices that are intersection lattices of real affine hyperplane arrangements is given by theorem [ t1 ] . for general geometric semilattices , a method for constructing bases which generalizes the broken circuit construction of @xcite is given by ziegler @xcite . this construction does not reduce to the construction given by theorem [ t1 ] in the case that the geometric semilattice is the intersection lattice of a real affine hyperplane arrangement . let @xmath3 be an essential arrangement of linear hyperplanes in @xmath4 . as before , let @xmath36 denote the set of intersections @xmath174 of subfamilies @xmath37 ( such intersections are necessarily nonempty in this case ) partially ordered by reverse inclusion . the finite lattice @xmath36 is called the _ intersection lattice _ of @xmath3 . it is a geometric lattice of length @xmath46 . now , let @xmath50 be an affine hyperplane in @xmath4 which is generic with respect to @xmath3 . genericity here means that @xmath202 for all @xmath203 . equivalently , @xmath204 and @xmath205 for all @xmath144-dimensional subspaces @xmath203 . let @xmath206 . this is an affine hyperplane arrangement induced in @xmath207 . we denote by @xmath208 its intersection semilattice . [ l3 ] @xmath209 . the top element @xmath64 of @xmath36 is the 0-dimensional subspace @xmath210 of @xmath4 . thus @xmath211 defines an order - preserving map @xmath212 , which is easily seen to be an isomorphism . the connected components of @xmath213 are pointed open convex polyhedral cones , that we call _ regions_. although none of these regions are bounded ( since @xmath3 is central ) , each region @xmath53 , nevertheless , induces a cycle @xmath56 in @xmath214 as follows . let @xmath215 denote the face lattice of the closed cone @xmath161 . that is , @xmath215 is the lower interval @xmath216 . clearly @xmath215 is isomorphic to the face lattice of the convex polytope @xmath217 , where @xmath92 is any affine hyperplane such that @xmath218 is nonempty and bounded . the map @xmath219 defined in section [ affine ] clearly embeds a copy of the dual of @xmath215 in @xmath36 . hence the image @xmath220 is a subposet of @xmath36 whose proper part is @xmath43-spherical ( meaning that its order complex is homeomorphic to @xmath221 ) . let @xmath56 be the fundamental cycle ( uniquely defined up to sign ) of the proper part of the subposet @xmath220 . [ t2 ] let @xmath3 be a central and essential hyperplane arrangement in @xmath4 and let @xmath50 be an affine hyperplane , generic with respect to @xmath3 . then the collection of cycles @xmath56 corresponding to regions @xmath53 such that @xmath222 is nonempty and bounded , form a basis of @xmath9 . this follows immediately from theorem [ t1 ] and the fact that @xmath223 ( lemma [ l3 ] ) . in order to apply theorem [ t2 ] to the examples given in subsequent sections we will need to choose an appropriate generic affine hyperplane and determine the regions whose affine slices are bounded . the following lemma provides a useful way of doing this . [ dot ] let @xmath3 be a central and essential hyperplane arrangement in @xmath4 . suppose @xmath224 is a nonzero element of @xmath4 such that the affine hyperplane @xmath225 through @xmath224 and normal to @xmath224 , is generic with respect to @xmath3 . then for any region @xmath53 of @xmath3 , @xmath226 is nonempty and bounded if and only if @xmath227 for all @xmath228 . @xmath229 suppose @xmath230 is nonempty and bounded . it is not difficult to see that if an affine slice of a cone is nonempty and bounded , then the cone is a cone over the affine slice . hence @xmath53 is a cone over @xmath231 . that is , every element of @xmath53 is a positive scalar multiple of an element of @xmath232 . it follows that since @xmath227 for all @xmath233 , @xmath234 for all @xmath235 . @xmath236 suppose @xmath237 is empty or unbounded . if the former holds then @xmath238 for all @xmath235 . indeed , if @xmath227 for some @xmath235 then @xmath239 we now assume @xmath232 is unbounded . then there is a sequence of points @xmath240 in @xmath231 whose distance from the origin goes to infinity . let @xmath241 be the unit vector in the direction of the vector @xmath242 . each @xmath241 is in the intersection of r and the unit sphere centered at the origin . hence , by passing to a subsequence if necessary , we can assume that the sequence of @xmath241 s converge to a unit vector @xmath243 in the closure of @xmath53 . since the cosine of the angle between @xmath241 and @xmath224 is @xmath244 , the cosine of the angles approach @xmath245 . hence the cosine of the angle between @xmath243 and @xmath224 is @xmath245 , or equivalently @xmath246 . since @xmath243 is in the closure of @xmath53 , either @xmath247 or there is a unique face @xmath54 of @xmath53 such that @xmath243 is in the interior of @xmath54 . if @xmath247 we are done . so suppose @xmath243 is in the interior of the face @xmath54 . if @xmath248 for all @xmath249 then the linear span of @xmath54 is an intersection of hyperplanes contained in the linear hyperplane with normal vector @xmath250 . this contradicts the genericity of @xmath225 . it follows that @xmath251 for some @xmath252 . if @xmath253 then there is a point @xmath254 that is close enough to @xmath255 so that @xmath256 and we are done . if @xmath234 then consider the point @xmath257 where @xmath258 . we have @xmath259 . by choosing @xmath260 to be small enough , we insure that the point @xmath261 is close enough to @xmath243 to be in @xmath54 , since @xmath243 is in the interior of @xmath54 . hence we have a point in @xmath54 whose dot product with @xmath224 is negative , putting us back in the previous case . theorem [ t2 ] can be extended to a geometric construction of bases for the whitney homology , or equivalently the orlik - solomon algebra , of the intersection lattice of a real central hyperplane arrangement . this involves the definition of a vector @xmath262 being _ totally generic _ with respect to the arrangement . since we will not pursue this direction we omit further mention of it . the arguments and results of the previous two sections can be generalized to oriented matroids . this generalization will be outlined in this section . the treatment here will be sketchy and can be skipped with no loss of continuity . the basics of oriented matroid theory will be assumed to be known . we refer to @xcite for all definitions and notation . let @xmath263 be an _ affine oriented matroid _ of rank @xmath91 and with _ affine face lattice _ @xmath264 , cf . * section 4.5 ) . the maximal elements of @xmath265 are the _ topes _ , corresponding to regions in the realizable case . let @xmath266 be the _ bounded complex _ ( a subcomplex of @xmath265 ) , and let @xmath267 be the set of bounded topes , i.e. , @xmath268 . we have from ( * ? ? ? 4.5.7 ) that @xmath265 is a shellable ball . furthermore , if @xmath269 then the order complex of the open interval @xmath270 in @xmath265 is homeomorphic to @xmath271 ( * ? ? ? therefore , each @xmath269 induces a spherical fundamental cycle @xmath272 in @xmath273 , where @xmath274 . [ p4 ] 1 . @xmath275 has the homotopy type of a wedge of @xmath276 copies of the @xmath277-sphere . @xmath278 is a basis of @xmath279 . the proof of proposition [ p2 ] generalizes . now , let @xmath280 be the _ intersection lattice _ ( or `` lattice of flats '' ) of the oriented matroid @xmath281 , and let @xmath282 be the `` zero map '' ( * ? ? ? furthermore , let @xmath283 $ ] . this is a geometric semilattice . the zero map restricts to an order - reversing surjection @xmath284 , and further to a surjection @xmath285 . [ p5 ] the map @xmath286 induces homotopy equivalence of order complexes . the proof of proposition [ p3 ] generalizes . here one uses that the quillen fibers @xmath287 , @xmath288 , are balls by ( * ? ? ? 4.5.7 ) , and hence contractible . the restriction of @xmath191 to an open interval @xmath289 in @xmath265 , with @xmath269 , gives an isomorphism of @xmath289 onto its image in @xmath290 . this image is a subposet of @xmath291 homeomorphic to the @xmath277-sphere . let @xmath292 be the corresponding fundamental cycle . [ t3 ] 1 . @xmath291 has the homotopy type of @xmath276 copies of the @xmath277-sphere . 2 . @xmath293 is a basis of @xmath294 . this follows from propositions [ p4 ] and [ p5 ] , since @xmath295 is an isomorphism and @xmath296 . the treatment of affine oriented matroids so far parallels that of affine hyperplane arrangements in section 3 . we will now move on to the oriented matroid version of the material in section 4 . let @xmath297 be an oriented matroid of rank @xmath91 , and let @xmath282 be the zero map to the corresponding intersection lattice @xmath280 . let @xmath298 be an extension of @xmath281 by a generic element @xmath299 . genericity here means that @xmath300 for every @xmath301 with @xmath302 , cf . consider the affine oriented matroid @xmath303 and let @xmath304 be its intersection semilattice . we have that @xmath305 and @xmath306 . [ l4 ] @xmath307 . this analog of lemma [ l3 ] is clear . it is basically a reformulation of the definition of genericity . let @xmath267 be the bounded topes of @xmath308 ( with respect to @xmath309 ) . because of the isomorphism @xmath310 we get cycles @xmath311 as before . [ t4 ] 1 . @xmath312 has the homotopy type of a wedge of @xmath276 copies of the @xmath277-sphere . @xmath293 is a basis for @xmath313 . this follows from theorem [ t3 ] and lemma [ l4 ] . the theorem gives a geometric method for constructing a basis for the homology of the geometric lattice of _ any _ orientable matroid . note that to define the set @xmath267 , and hence the basis , we must make a generic extension of @xmath281 . different extensions will yield different bases . the hyperplane arrangement @xmath314 in @xmath315 , where @xmath316 , is known as the _ braid arrangement _ or the _ type @xmath0 coxeter arrangement_. the orthogonal reflection @xmath317 across the hyperplane @xmath318 acts on @xmath319 by switching its @xmath84th and @xmath320th coordinates . these reflections generate the symmetric group @xmath11 acting on @xmath315 by permuting coordinates . the braid arrangement is not essential . to make it essential let @xmath321 and define @xmath322 then @xmath323 is an essential central hyperplane arrangement in the @xmath324-dimensional space @xmath325 . it is clear that the intersection lattices @xmath326 and @xmath327 are isomorphic . they are also isomorphic to the partition lattice @xmath10 . indeed , for each @xmath328 , let @xmath329 be the linear subspace of @xmath315 consisting of all points @xmath330 such that @xmath331 whenever @xmath84 and @xmath320 are in the same block of @xmath332 . the map @xmath333 is an isomorphism from @xmath10 to @xmath327 . let @xmath334 denote the inverse of this isomorphism . the arrangement @xmath323 has @xmath335 regions which are all simplicial cones and are in a natural one - to - one correspondence with the elements of the associated coxeter group @xmath11 . under this correspondence a permutation @xmath336 corresponds to the region @xmath337 consider the cycle @xmath338 whose general construction was given in section [ cent ] . we now give a simple explicit description of the image of @xmath339 in @xmath340 under the isomorphism @xmath334 . to split a permutation @xmath13 at positions @xmath341 in @xmath342 $ ] is to form the partition with @xmath343 blocks , @xmath344 ( to split @xmath15 at the empty set of positions is to form the partition with one block . ) let @xmath14 denote the induced subposet of @xmath10 consisting of all partitions obtained by splitting the permutation @xmath15 . clearly @xmath14 is isomorphic to the lattice of subsets of @xmath342 $ ] . hence @xmath345 is spherical . [ funda ] for all @xmath13 , the image @xmath346 is the fundamental cycle of @xmath345 . recall the map @xmath347 that takes cells to their affine span ( defined in section [ affine ] ) . we show that @xmath334 restricts to an isomorphism from the subposet @xmath348 of @xmath349 to @xmath14 . observe that elements of the face lattice @xmath350 are sets of the form @xmath351 where @xmath352 . the linear span of such a set is the subspace @xmath353 hence @xmath348 is the poset of subspaces of the form given in ( [ space ] ) ordered by reverse inclusion . clearly @xmath334 takes the subspace given in ( [ space ] ) to the partition given in ( [ split ] ) . we now choose a vector @xmath354 in @xmath325 that satisfies the hypothesis of lemma [ dot ] and use lemma [ dot ] to describe the permutations @xmath13 for which the regions @xmath355 are bounded . [ bounda ] let @xmath356 . then the affine hyperplane @xmath357 is generic with respect to the arrangement @xmath323 of @xmath325 . moreover , for all @xmath13 , @xmath355 is bounded if and only if @xmath358 . recall that genericity is equivalent to the condition that for all @xmath144-dimensional subspaces @xmath359 , @xmath360 . the @xmath144-dimensional intersections of hyperplanes in @xmath323 have the form @xmath361 where @xmath362 and @xmath363 is a partition of @xmath364 $ ] . let @xmath365 be defined by @xmath366 and @xmath367 . one can easily check that @xmath368 . hence @xmath357 is generic . to prove that @xmath369 is bounded if and only if @xmath370 , we apply lemma [ dot ] . note that for all @xmath371 , @xmath372 suppose @xmath370 . for all @xmath373 , we have @xmath374 and @xmath375 . hence @xmath376 . it follows that @xmath377 for all @xmath378 . by lemma [ dot ] @xmath369 is bounded . now suppose @xmath379 . clearly there exists an @xmath373 such that @xmath380 for all @xmath381 . for such an @xmath255 , we have @xmath382 . we conclude by lemma [ dot ] that @xmath369 is not bounded . [ splitting basis @xcite ] [ splita ] for @xmath13 , let @xmath25 be the fundamental cycle of @xmath383 . then @xmath384 is a basis for @xmath385 . this follows from theorem [ t2 ] and propositions [ funda ] and [ bounda ] . the type @xmath1 coxeter arrangement is the hyperplane arrangement @xmath386 the orthogonal reflections across the hyperplanes generate the hyperoctahedral group @xmath29 . we will view the elements of @xmath29 as signed permutations , that is , words consisting of @xmath26 distinct letters from @xmath364 $ ] where any of the letters can have a bar placed above it . it will also be convenient to express elements of @xmath29 as pairs @xmath387 where @xmath13 and @xmath388 . if @xmath389 then @xmath390 does not have a bar over it and if @xmath391 then @xmath390 has a bar over it . for example , the signed permutation @xmath392 can be expressed as @xmath393 a signed permutation @xmath387 maps @xmath394 to @xmath395 . the arrangement @xmath396 is essential and has @xmath397 regions which are all simplicial cones and are in a natural one - to - one correspondence with the elements of the hyperoctahedral group @xmath29 . under this correspondence a signed permutation @xmath398 corresponds to the region @xmath399 the intersection lattice @xmath400 is isomorphic to the signed partition lattice @xmath401 which is defined as follows . let @xmath332 be a partition of the set @xmath28 . the block containing @xmath245 is called the _ zero block_. to _ bar _ an element of a block of @xmath332 is to place a bar above the element and to unbar a barred element is to remove the bar . a signed partition is a partition of the set @xmath28 in which any of the nonminimal elements of any of the nonzero blocks are barred . for example , @xmath402 is a signed partition of @xmath403 . it will be convenient to sometimes express a barred letter @xmath404 of a signed partition as @xmath405 and an unbarred letter as @xmath406 . to bar a block @xmath407 in a signed partition is to bar all unbarred elements in @xmath407 and to unbar all barred elements in @xmath407 . we denote this by @xmath408 . to unbar a block @xmath407 is to unbar all barred elements of @xmath407 . we denote this by @xmath409 . for example , @xmath410 let @xmath401 be the poset of signed partitions of @xmath28 with order relation defined by @xmath411 if for each block @xmath407 of @xmath332 , either @xmath407 is contained in a nonzero block of @xmath412 , @xmath408 is contained in a nonzero block of @xmath412 or @xmath409 is contained in the zero block of @xmath412 . for example @xmath413 @xmath414 and @xmath415 the poset @xmath27 is an example of a dowling lattice @xcite . for each signed partition @xmath416 , let @xmath329 be the linear subspace of @xmath315 consisting of all points @xmath417 such that * @xmath331 whenever @xmath84 and @xmath320 are in the same block of @xmath332 and both are barred or both are unbarred , * @xmath418 whenever @xmath84 and @xmath320 are in the same block of @xmath332 and one is barred and the other is unbarred , * @xmath419 whenever @xmath84 is in the zero block . the map @xmath420 is an isomorphism from @xmath401 to @xmath421 . let @xmath334 be the inverse of this isomorphism . the cycle @xmath422 maps under @xmath334 to the fundamental cycle of a spherical subposet of @xmath401 , which we now describe . to split a signed permutation @xmath387 at positions @xmath341 in @xmath423 is to form the signed partition with blocks @xmath424 where @xmath425 ( here @xmath426 ) , and for all @xmath427 , @xmath428 is either @xmath429 or @xmath430 ( here @xmath431 ) . for example , if we split the signed permutation @xmath432 at positions @xmath433 we get the signed partition @xmath434 for each signed permutation @xmath435 , let @xmath436 be the induced subposet of @xmath27 consisting of all signed partitions obtained by splitting the signed permutation @xmath387 . just as for type a , @xmath437 is spherical because @xmath436 is isomorphic to the lattice of subsets of @xmath364 $ ] . [ fundb ] for all @xmath435 , the image @xmath438 is the fundamental cycle of @xmath437 . let @xmath439 be the map that takes cells to their affine span ( cf . section [ affine ] ) . we show that @xmath334 restricts to an isomorphism from the subposet @xmath440 of @xmath400 to @xmath436 . observe that elements of the face lattice @xmath441 are sets of the form @xmath442 where @xmath443 . the linear span of such a set is the subspace @xmath444 hence @xmath440 is the poset of subspaces of the form given in ( [ space2 ] ) ordered by reverse inclusion . clearly @xmath334 takes the subspace given in ( [ space2 ] ) to the signed partition obtained by splitting @xmath387 at positions @xmath445 . let @xmath398 . we say that @xmath390 is a _ right - to - left maximum _ of @xmath387 if @xmath446 for all @xmath447 . [ bounded ] let @xmath448 . then the affine hyperplane @xmath449 is generic with respect to the arrangement @xmath396 . moreover , for all @xmath398 , @xmath450 is bounded if and only if all right - to - left maxima of @xmath387 are unbarred . the @xmath144-dimensional intersections of hyperplanes in @xmath396 have the form @xmath451 where @xmath452 , @xmath453 is a partition of @xmath28 and @xmath454 , for @xmath455 . let @xmath456 be defined by @xmath457 one can easily check that @xmath458 . hence @xmath360 for all @xmath459dimensional spaces @xmath460 in @xmath400 , which means that @xmath449 is generic . now suppose @xmath435 and @xmath461 is bounded . assume that @xmath462 is a right - to - left maximum that is barred . we will reach a contradiction of lemma [ dot ] by producing a vector @xmath463 in the closure of @xmath461 such that @xmath464 . let @xmath465 be such that @xmath466 clearly @xmath467 . we have @xmath468 since @xmath469 for all @xmath470 , we have @xmath471 combining this with the previous inequality yields @xmath472 . now suppose @xmath435 and all right - to - left maxima of @xmath387 are unbarred . let @xmath473 be the positions of the right - to - left maxima . . then @xmath475 where @xmath476 . we have @xmath477 for all @xmath478 , since @xmath479 is a right - to - left maximum and @xmath390 is not . also , @xmath480 for all @xmath478 , since @xmath479 is unbarred . thus we have shown that @xmath482 . by lemma [ dot ] , @xmath483 is bounded . [ splitb ] for @xmath435 , let @xmath484 be the fundamental cycle of the spherical poset @xmath437 . then @xmath485 is a basis for @xmath486 . this follows from theorem [ t2 ] and propositions [ fundb ] and [ bounded ] . [ dow ] the rank of @xmath487 is @xmath488 one can construct a signed permutation by inserting ( barred or unbarred ) elements @xmath489 one at a time from largest to smallest . to construct a signed permutation in which the right - to - left maxima are unbarred , first we insert @xmath26 into the empty word . there is only one way to do this since @xmath26 must be unbarred . now suppose we have inserted @xmath490 . if we insert @xmath491 at the end of the partially constructed word , @xmath491 will be a right - to - left maximum in the final word and must be unbarred . if we insert @xmath491 in any of the other @xmath320 positions @xmath491 will not be a right - to - left maximum and can therefore be barred or unbarred . hence there are @xmath492 ways to insert @xmath491 , for each @xmath493 . [ dow ] the posets @xmath10 and @xmath27 are examples of dowling lattices . gottlieb and wachs ( * ? ? ? * section 9 ) have constructed splitting bases for general dowling lattices . the splitting basis for @xmath10 is a special case of this dowling lattice construction . for @xmath494 , the dowling lattice construction produces the basis consisting of cycles @xmath484 where @xmath435 is such that all _ left - to - right _ ( rather than right - to - left ) maxima of @xmath387 are unbarred . although this basis is similar in appearance to the type @xmath1 splitting basis given here , one can show that it can not be obtained from our geometric construction . ( consider the intersection of @xmath60 with the cube as in figure 4 . the union of the closure of the regions corresponding to the signed permutations whose left - to - right maxima are unbarred , is not simply connected . hence these regions can not be the bounded regions of a generic slice . ) however by using techniques similar to that of @xcite , one can get a variation of the dowling lattice splitting basis which does reduce to the type @xmath1 splitting basis given here . although the only dowling lattices that are intersection lattices of real hyperplane arrangements are the partition lattice and the signed partition lattice , there are other dowling lattices that are intersection lattices of complex hyperplane arrangements ( cf . * section 8) ) . it would be interesting to find a geometric interpretation of the above mentioned variation of the splitting basis for such dowling lattices . the splitting basis of type @xmath0 is used in @xcite and @xcite to obtain information about the representation of the symmetric group on the homology of the partition lattice . unfortunately , the splitting basis of type @xmath1 ( or @xmath2 ) does not appear to reveal much about the representation of the coxeter group on homology . see @xcite , @xcite and @xcite for work on the representation of @xmath29 on the homology of the signed partition lattice . the type @xmath2 coxeter arrangement is the hyperplane arrangement @xmath495 the orthogonal reflections across the hyperplanes generate the coxeter group @xmath496 , which is the subgroup of @xmath29 consisting of all signed permutations that have an even number of bars . clearly @xmath497 is a subarrangement of @xmath396 and its intersection lattice is isomorphic to @xmath498 , the join - sublattice of @xmath499 consisting of all signed partitions whose zero block does not have size @xmath500 . this isomorphism is denoted by @xmath501 and is the restriction of the isomorphism @xmath502 defined in the previous section . the arrangement @xmath497 is essential and has @xmath503 regions which are simplicial cones in one - to - one correspondence with the elements of @xmath496 . under this correspondence the signed permutation @xmath387 corresponds to the region @xmath504 note that the hyperplane @xmath505 divides the region @xmath506 into the regions @xmath461 and @xmath507 , where @xmath508 for @xmath509 and @xmath510 . for each signed permutation @xmath511 , let @xmath512 be the induced subposet of @xmath513 consisting of all signed partitions obtained by splitting either the signed permutation @xmath387 or the signed permutation @xmath514 at all positions in a subset of @xmath423 whose smallest element is not @xmath144 . although it is not as evident as for types @xmath0 and @xmath1 , this induced subposet is also a subset lattice . the induced subposet @xmath515 of @xmath498 is isomorphic to the lattice of subsets of @xmath364 $ ] . define the map @xmath516 by letting @xmath517 be the signed partition obtained by splitting @xmath387 at all positions in @xmath518 if @xmath144 is not the smallest element of @xmath518 , and by splitting @xmath514 at all positions in @xmath519 otherwise . we leave it to the reader to check that this map is an isomorphism from the dual of @xmath520 to @xmath521 . [ fundd ] for all @xmath511 , the image @xmath522 is the fundamental cycle of @xmath523 . we show that @xmath334 restricts to an isomorphism from the subposet @xmath524 of @xmath525 to @xmath526 . the elements of @xmath527 are sets that have one of the following forms : @xmath528 @xmath529 @xmath530 @xmath531 where @xmath532 . the linear span of the cells are the respective subspaces : @xmath533 @xmath534 @xmath535 @xmath536 where @xmath537 . clearly the images of such subspaces under @xmath334 are precisely the elements of @xmath521 . [ boundd ] let @xmath538 . then the affine hyperplane @xmath449 is generic with respect to the arrangement @xmath497 . moreover , for all @xmath539 , @xmath540 is bounded if and only if @xmath541 and all right - to - left maxima of @xmath387 are unbarred . since by proposition [ bounded ] , @xmath449 is generic with respect to the arrangement @xmath396 , it is generic with respect to any subarrangement of @xmath396 ; in particular it is generic with respect to @xmath497 . since the hyperplane @xmath542 divides @xmath506 into the regions @xmath543 and @xmath544 , the region @xmath545 is bounded if and only if both @xmath546 and @xmath547 are bounded . by proposition [ bounded ] , both regions are bounded if and only if all right - to - left maxima of both @xmath387 and @xmath514 are unbarred . this happens if and only if @xmath548 is not a right - to - left maximum , i.e. , @xmath541 , and all right - to - left maxima of @xmath387 are unbarred . [ splitd ] for each @xmath511 , let @xmath549 be the fundamental cycle of @xmath523 . then @xmath550 @xmath551 is a basis for @xmath552 . this follows from theorem [ t2 ] and propositions [ fundd ] and [ boundd ] . the following corollary is well - known in the theory of coxeter arrangements . the rank of @xmath553 is @xmath554 we construct a signed permutation @xmath511 with all right - to - left maxima unbarred and @xmath541 by first choosing @xmath548 in @xmath555 ways and then choosing the signed permutation @xmath556 so that every right - to - left maximum is unbarred . by the proof of corollary [ dow ] , this signed permutation on @xmath555 letters can be chosen in @xmath557 ways . if the number of bars in @xmath556 is even then @xmath548 must be unbarred ; otherwise @xmath548 must be barred . we now consider a family of posets which interpolates between the type d partition lattice and the type b partition lattice . for @xmath558 $ ] , let @xmath559 be the join - sublattice of @xmath499 consisting of all signed partitions whose zero block is not @xmath560 for @xmath561\setminus t$ ] . clearly @xmath562 and @xmath563 ) = \pi^b_n$ ] . the lattice @xmath559 is the intersection lattice of the hyperplane arrangement @xmath564 these interpolating arrangements were introduced by zaslavsky @xcite . the following theorem generalizes theorems [ splitb ] and [ splitd ] . there are two types of regions of the hyperplane arrangement @xmath569 , namely @xmath461 for @xmath570 and @xmath571 , and @xmath506 for @xmath572 and @xmath573 . by propositions [ fundb ] and [ fundd ] , @xmath484 and @xmath574 are the respective images ( up to sign ) of @xmath575 and @xmath576 , under the isomorphism @xmath334 . we proceed as in the proof of theorem [ splitd ] . for @xmath577 , @xmath449 is generic with respect to the arrangement @xmath569 . hence the result now follows from theorem [ t2 ] , propositions [ bounded ] and [ boundd ] . . then the number of signed permutations @xmath435 such that @xmath571 and all right - to - left maxima are unbarred is @xmath581 . the number of signed permutations @xmath511 such that @xmath582 and all right - to - left maxima are unbarred is @xmath583 . now suppose @xmath584 . then the number of signed permutations @xmath435 such that @xmath571 and all right - to - left maxima are unbarred is @xmath585 . the number of signed permutations @xmath511 such that @xmath586 and all right - to - left maxima are unbarred is @xmath587 . in either case , the total number of elements in the basis is @xmath588 . jzefiak and sagan @xcite have studied other families of hyperplane arrangements which interpolate between coxeter arrangements . one can apply our results to these arrangements . the family of arrangements that interpolate between @xmath589 and @xmath590 are particularly amenable to our approach . for @xmath591 $ ] , let @xmath592 be the arrangement in @xmath315 , @xmath593 let @xmath594 be the induced subposet of @xmath10 consisting of all partitions @xmath332 such that the block of @xmath332 containing @xmath26 is either a singleton or has nonempty intersection with @xmath595 . this is the intersection lattice of @xmath592 . the following theorem generalizes theorem [ splita ] by providing a splitting basis for the homology of @xmath594 . recall that in section [ braid ] we defined @xmath25 to be the fundamental cycle of @xmath596 , for each @xmath597 . there are easier and more direct ways to prove theorem [ at ] . for instance , one can restrict the second el - labeling for @xmath10 given in the proof of theorem 6.2 of @xcite , to @xmath594 . the induced shelling basis is precisely the basis given in theorem [ at ] . in @xcite , stanley showed that the restriction to @xmath603 of the representation of @xmath11 on @xmath604 is the regular representation . in @xcite , it was observed that the splitting basis for @xmath605 makes this fact transparent . indeed , the permutations that fix @xmath26 permute the basis cycles @xmath25 . a similar phenomenon occurs for @xmath594 if @xmath13 satisfies @xmath370 and @xmath610 then so does @xmath611 for all @xmath612\setminus t } \times s_{\{n\}}$ ] . in fact , the elements of @xmath595 occupy the same set of positions in @xmath613 as in @xmath614 . since @xmath615 , we see that @xmath607\setminus t } $ ] acts on @xmath608 by permuting basis cycles . also , the orbit of @xmath25 is determined by the set of positions that elements of @xmath595 occupy in @xmath15 . hence the number of orbits is @xmath609 . john shareshian [ personal communication ] has found an alternative proof of corollary [ reg ] which involves a computation of the mbius function of the @xmath180-invariant subposet of @xmath594 for @xmath612\setminus t}$ ] by means of crapo s complementation formula @xcite . another class of partition posets with a splitting basis is the class of @xmath46-divisible partition lattices @xcite , or the more general restricted block size partition lattices considered in @xcite and @xcite . these are not geometric lattices in general ; but they are intersection lattices of subspace arrangements and they are shellable . the cycles in the basis are polytopal . the symmetric group acts on these lattices and the splitting basis reveals much information about the representation of the symmetric group on homology . it would be interesting to find a geometric explanation for this spitting basis . one might also consider restricted block size partition subposets of @xmath594 .

we use the theory of hyperplane arrangements to construct natural bases for the homology of partition lattices of types @xmath0 , @xmath1 and @xmath2 . this extends and explains the `` splitting basis '' for the homology of the partition lattice given in @xcite , thus answering a question asked by r. stanley . more explicitly , the following general technique is presented and utilized . let @xmath3 be a central and essential hyperplane arrangement in @xmath4 . let @xmath5 be the bounded regions of a generic hyperplane section of @xmath3 . we show that there are induced polytopal cycles @xmath6 in the homology of the proper part @xmath7 of the intersection lattice such that @xmath8 is a basis for @xmath9 . this geometric method for constructing combinatorial homology bases is applied to the coxeter arrangements of types @xmath0 , @xmath1 and @xmath2 , and to some interpolating arrangements . _ dedicated to richard stanley on the occasion of his 60th birthday _

the second law of newtonian mechanics states that if @xmath0 is the force acting on a point particle of mass @xmath1 and @xmath2 is its acceleration , then @xmath3 . in a sense , the physical meaning of this expression lies in its tacit assumptions , namely that forces are vectors , that is , elements of a vector space , and as such they sum . this experimental fact embodied in the second law is what prevent us from considering the previous identity as a mere definition of _ force_. coming to the study of the rigid body , one can deduce the first cardinal equation of mechanics @xmath4 , where @xmath5 is the affine point of the center of mass , @xmath6 is the total mass and @xmath7 is the resultant of the external applied forces . this equation does not fix the dynamical evolution of the body , indeed one need to add the second cardinal equation of mechanics @xmath8 , where @xmath9 and @xmath10 , are respectively , the total angular momentum and the total mechanical momentum with respect to an arbitrary fixed point @xmath11 . naively adding the applied forces might result in an incorrect calculation of @xmath10 . as it is well known , one must take into account the line of action of each force @xmath12 in order to determine the _ central axis _ , namely the locus of allowed application points of the resultant . these considerations show that applied forces do not really form a vector space . this unfortunate circumstance can be amended considering , in place of the force , the field of mechanical momenta that it determines ( the so called _ dynamical screw _ ) . these type of fields are constrained by the law which establishes the change of the mechanical momenta under change of pole @xmath13 an analogy between momenta and velocities , and between force resultant and angular velocity is apparent considering the so called _ fundamental formula of the rigid body _ , namely a constraint which characterizes the velocity vector field of the rigid body @xmath14 the correspondence can be pushed forward for instance by noting that the concept of _ instantaneous axis of rotation _ is analogous to that of _ central axis_. screw theory explores these analogies in a systematic way and relates them to the lie group of rigid motions on the euclidean space . perhaps , one of the most interesting consequences of screw theory is that it allows us to fully understand that angular velocities should be treated as vectors applied to the instantaneous axis of rotation , rather than as free vectors . this fact is not at al obvious . let us recall that the angular velocity is defined through poisson theorem , which states that , given a frame @xmath15 moving with respect to an absolute frame @xmath16 , any normalized vector @xmath17 which is fixed with respect to @xmath15 satisfies @xmath18 in the original frame @xmath16 , where @xmath19 is unique . the uniqueness allows us to unambiguously define @xmath19 as the angular velocity of @xmath15 with respect to @xmath16 . as the vectors @xmath17 are free , their application point is not fixed and so , according to this traditional definition , @xmath19 is not given an application point . this fact seems close to intuition . indeed , let us consider foucault s 1851 famous experiment performed at the paris observatory . by using a pendulum he was able to prove that the earth rotates with an angular velocity which coincides with that inferred from the observation of distant stars . of course , the choice of paris was not essential , and the measurement would have returned the same value for the angular velocity were it performed in any other place on earth . in fact , the reason for assigning to the angular velocity an application point in the instantaneous axis of rotation becomes clear only in very special applications , and in particular when the composition of rigid motions is considered . this fact will be fully justified in section [ njc ] . here we just wish to illustrate how , using the analogy between forces and angular velocities , it is possible to solve non - trivial problems on the composition of motions . consider , for instance , four frames @xmath20 , @xmath21 , where @xmath22 is the absolute frame and @xmath23 , @xmath24 , moves with respect to @xmath20 with an angular velocity @xmath25 . let us suppose that at the given instant of interest , and for @xmath24 , the instantaneous axes of rotation of @xmath23 as seen from @xmath20 , lie all in the same plane as shown in figure [ fun ] . we can apply the well known rules of statics , for instance using the funicular polygon @xcite , to obtain the angular velocity @xmath26 and the instantaneous axis of rotation of @xmath27 with respect to @xmath22 . motion with respect to @xmath22 by using the funicular polygon method . this method was originally developed for finding the central axis in problems of statics.,width=340 ] it is also interesting to observe if a frame @xmath28 rotates with angular velocity @xmath19 with respect to @xmath29 , and @xmath29 rotates with angular velocity @xmath30 with respect to @xmath22 , and if the two instantaneous axes of rotation are parallel and separated by an arm of length @xmath31 , then , at the given instant , @xmath28 translates with velocity @xmath32 in a direction perpendicular to the plane determined by the two axes . as a consequence , any act of translation can be reduced to a composition of acts of pure rotation . this result is analogous to the usual observation that two opposite forces @xmath0 and @xmath33 with arm @xmath31 generate a constant mechanical momenta of magnitude @xmath34 and direction perpendicular to the plane determined by the two forces . as a consequence , any applied momenta can be seen as the effect of a couple of forces . of course , screw theory has other interesting consequences and advantages . we invite the reader to discover and explore them in the following sections . the key ideas leading to screw theory included in this article have been taught at a second year undergraduate course of `` rational mechanics '' at the faculty of engineer of florence university ( saved for the last technical section ) . we shaped this text so as to be used by our students for self study and by any other scholars who might want to introduce screw theory in an undergraduate course . indeed , we believe that it is time to introduce this beautiful approach to mechanics already at the level of undergraduate university programs . screw theory is venerable ( for an account of the early history see @xcite ) . it originated from the works of euler , mozzi and chasles , who discovered that any rigid motion can be obtained as a rotation followed by a translation in the direction of the rotation s axis ( this is the celebrated chasles s theorem which was actually first obtained by giulio mozzi @xcite ) , and by poinsot , chasles and mbius , who noted the analogy between forces and angular velocities for the first time @xcite . it was developed and reviewed by sir r. ball in his 1870 treatise @xcite , and further developed , especially in connection with its algebraic formulation , by clifford , kotelnikov , zeylinger , study and others . unfortunately , by the end of the nineteenth century it was essentially forgotten to be then fully rediscovered only in the second half of the twentieth century . it remains largely unknown and keeps being rediscovered by various authors interested in rigid body mechanics ( including this author ) . unfortunately , screw theory is usually explained following descriptive definitions rather than short axiomatic lines of reasoning . as a result , the available introductions are somewhat unsatisfactory to the modern mathematical and physical minded reader . perhaps for this reason , some authors among the few that are aware of the existence of screw theory claim that it is too complicated to deserve to be taught . for instance , the last edition of goldstein s textbook @xcite includes a footnote which , after introducing the full version of chasles theorem ( sect . [ mvj ] ) , which might be regarded as the starting point of screw theory , comments _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [ @xmath35 there seems to be little present use for this version of chasles theorem , nor for the elaborate mathematics of screw motions elaborated at the end of the nineteenth century . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ were it written in the fifties of the last century this claim could have been shared , but further on screw theory has become a main tool for robotics @xcite where it is ordinarily used . furthermore , while elaborate the mathematics of screw theory simplifies the development of mechanics . admittedly , however , some people could be dissatisfied with available treatments and so its main advantages can be underestimated . we offer here a shorter introduction which , hopefully , could convince these readers of taking a route into screw theory . let us comment on some definitions of screw that can be found in the literature , so as to justify our choices . a first approach , that this author does not find appealing , introduces the screw by means of the concept of _ motor_. this formalism depends on the point of reduction , and one finds the added difficulty of proving the independence of the various deductions from the chosen reduction point . it hides the geometrical content of the screw and makes proofs lengthier . nevertheless , it must be said that the motor approach could be convenient for reducing screw calculations to a matter of algebra ( the so called screw calculus @xcite ) . in a similar vein , some references , including selig s @xcite , introduce the screw from a matrix formulation that tacitly assumes that a choice of reference frame has been made ( thus losing the invariance at sight of the definition ) . still concerning the screw definition , some literature follows the practical and traditional approach which introduces the screw from its properties ( screw axis , pitch , etc . ) @xcite , like in old fashioned linear algebra where one would have defined a vector from its direction , verse and module , instead of defining it as an element of a vector space ( to complicate matters , some authors define the screw up to a positive constant , in other words they work with a projective space rather that a vector space ) . this approach could be more intuitive but might also give a false confidence of understanding , and it is less suited for a formal development of the theory . it is clear that the vector space approach in linear algebra , while less intuitive at first , proves to be much more powerful than any descriptive approach . of course , one has to complement it with the descriptive point of view in order to help the intuition . in my opinion the same type of strategy should be followed in screw theory , with a maybe more formal introduction , giving a solid basis , aided by examples to help the intuition . since descriptive intuitive approaches are not lacking in the literature , this work aims at giving a short introduction of more abstract and geometrical type . it should be said that at places there is an excess of formality in the available presentations of screw theory . i refer to the tendency of giving separated definitions of screws , one for the kinematical _ twist _ describing the velocity field of the rigid body , and the other for the _ wrench _ describing the forces acting on the body . this type of approach , requiring definitions for screws and their dual elements ( sometimes called co - screws ) , lengthens the presentation and forces the introduction and use of the dual space of a vector space , a choice which is not so popular especially for undergraduate teaching . who adopts this point of view argues that it should also be adopted for forces in mechanics , which should be treated as 1-forms instead as vectors . this suggestion , inspired by the concept of conjugate momenta of lagrangian and hamiltonian theory , sounds more modern , but would be geometrically well founded only if one could develop mechanics without any mention to the scalar product . the scalar product allows us to identify a vector space with its dual and hence to work only with the former . if what really matters is the pairing between a vector space and its dual then , as this makes sense even without scalar product , we could dispense of it . it is easy to realize that in order to develop mechanics we need a vector space ( and/or its dual ) as well as a scalar product and an orientation ( although most physical combinations of interest might be rewritten so as a to get rid of it , e.g. the kinetic energy is @xmath36 $ ] ) . analogously , in screw theory , it could seem more appealing to look at kinematical twists as screws and to dynamical wrenches as co - screws , but geometrically this choice does not seem compelling , and in fact it is questionable , given the price to be paid in terms of length and loss of unity of the presentation . therefore , we are going to use just one mathematical entity - the screw - emphasizing the role of the screw scalar product in identifying screws and dual elements . in this work i took care at introducing screw theory in a way as far as coordinate independent as possible , but avoiding the traditional descriptive route . in this approach the relation with the lie algebra of rigid maps becomes particularly transparent . finally , most approaches postpone the definition of screw after the examples of systems of applied forces from which the idea of screw can been derived . i think that it is better to introduce the screw first and then to look at the applications . in this way , through some key choices , i have obtained a hopefully clear and straightforward introduction to screw theory , which is at the same time mathematically rigorous . my hope is that after reading these notes , the reader will share the author s opinion that screw theory is indeed `` the right '' way of teaching rigid body mechanics as the tight relation with the lie group of rigid maps suggests . in this section we define the _ screw _ and prove some fundamental properties . specific applications will appear in the next sections . let us denote with @xmath37 the affine euclidean space modeled on the three dimensional vector space @xmath38 . the space @xmath38 is endowed with a positive definite scalar product @xmath39 , and is given an orientation . this structure is represented with a triple @xmath40 where @xmath41 denotes the orientation . note that thanks to this structure a vector product @xmath42 can be naturally defined on @xmath38 . points of @xmath37 are denoted with capital letters e.g. @xmath43 while points in @xmath38 are denoted as @xmath44 we shall repeatedly use the fact that the mixed product @xmath45 changes sign under odd permutations of its terms and remains the same under even permutations . a vector field is a map @xmath46 . an _ applied vector _ is an element of @xmath47 , namely a pair @xmath48 where @xmath49 is the _ application point _ of @xmath48 . a _ sliding vector _ is an equivalence class of applied vectors , where two applied vectors @xmath48 and @xmath50 are equivalent if @xmath51 and for some @xmath52 , @xmath53 , namely they have the same line of action . we shall preferably use the concept of applied vector even in those cases in which it could be equivalently replaced by that of sliding vector . the reason is that the concept of sliding vector is superfluous because it is more convenient to regard applied and sliding vectors as special types of screws . occasionally , we shall use the concept of _ reference frame _ which is defined by a choice of _ origin _ @xmath54 , and of positive oriented orthonormal base @xmath55 for @xmath40 . once a reference frame has been fixed , any point @xmath56 is univocally determined by its coordinates @xmath57 , defined through the equation @xmath58 . in order to lighten the formalism we shall consider different physical vector quantities , such as position , velocity , linear momenta , force , mechanical momenta , as elements of the same vector space @xmath38 . a more rigorous treatment would introduce a different vector space for each one of these concepts . the reader might imagine to have fixed the dimension units . it is understood that , say , a linear momenta can not be summed to a force even though in our treatment they appear to belong to the same vector space . a _ screw _ is a vector field @xmath59 which admits some @xmath60 in such a way that for any two points @xmath61 @xmath62 for any screw @xmath63 the vector @xmath64 is unique , indeed if @xmath64 and @xmath65 satisfy the above equation , then subtracting the corresponding equations @xmath66 and from the arbitrariness of @xmath67 , @xmath68 . the vector @xmath64 is called the _ resultant of the screw_. if the resultant of the screw vanishes then @xmath69 does not depend on @xmath67 and the screw is said to be _ constant_. equation ( [ nuv ] ) is the _ constitutive equation of the screw_. if @xmath63 is a screw the quantity @xmath70 does not depend on the point and in called the _ scalar invariant of the screw_. the _ vector invariant of the screw _ is the quantity ( independent of @xmath67 ) and defined by @xmath71 thus if @xmath72 the vector invariant of the screw is the projection of @xmath69 on the direction given by the resultant , and it is actually independent of @xmath67 . the screws form a vector space @xmath73 and the map which sends @xmath63 to @xmath64 is linear . if @xmath74 and @xmath75 are two screws @xmath76 multiplying by @xmath77 the first equation and adding the latter multiplied by @xmath78 we get @xmath79 which implies that the screws form a vector space and that the resultant of the screw @xmath80 is @xmath81 , that is , the map @xmath82 is linear . given two screws @xmath74 and @xmath75 let us consider the quantity @xmath83 for any two points @xmath61 , @xmath84 . by definition @xmath85 and @xmath86 , thus @xmath87+{\bm{s}}_2 \cdot[{\bm{s}}_1 \times ( p - q)]\}\\ & = { \bm{s}}_1 \cdot s_2(q)+ { \bm{s}}_2 \cdot s_1(q).\end{aligned}\ ] ] according to the previous result we can simply write @xmath88 in place of @xmath89 . the _ screw scalar product _ is the symmetric bilinear map @xmath90 which sends @xmath91 to @xmath88 . note that the scalar invariant of a screw is one - half the screw scalar product of the screw by itself . since this scalar invariant can be negative , the screw scalar product on @xmath73 is not positive definite . nevertheless , we shall see that it is non - degenerate ( sect . [ gvo ] ) . the cartesian product @xmath92 endowed with the usual sum and product by scalar gives the direct sum @xmath93 . typically , there will be three ways to construct screws out of ( applied ) vectors . the easy proofs to the next two propositions are left to the reader . the map @xmath94 given by @xmath95 sends a ( free ) vector to a constant screw . the map @xmath96 given by @xmath97 sends an applied vector to a screw . the map @xmath98 given by @xmath99 sends a pair given by an applied vector and a free vector to a screw . the screws in the image of @xmath77 will be called constant or _ free screws_. the screws in the image of @xmath78 will be called _ applied screws_. clearly , by the constitutive equation of the screw , the map @xmath100 is surjective . in particular , every screw is the sum of a free screw and an applied screw . let @xmath101 , then this linear map is bijective . its inverse @xmath102 is called _ motor reduction _ at @xmath11 . once we agree on the reduction point @xmath11 , any pair @xmath103 as in the previous proposition is called a _ motor _ at @xmath11 . sometimes we shall write @xmath104 for @xmath105 , thus the motor at @xmath11 reads @xmath106 . often , for reasons that will be soon clear , we will prefer to represent the ordered pair in a column form of two elements of @xmath38 . we can write the found bijective correspondence between @xmath73 and @xmath107 as follows @xmath108 in this representation the screw scalar product is given by @xmath109 , thus is is mediated by the matrix @xmath110 where @xmath111 is the identity map . let us now recall that any point @xmath54 can be used as _ origin _ , namely it allows us to establish a bijective correspondence between @xmath37 and @xmath38 given by @xmath112 . if we additionally introduce a positive oriented orthonormal base then we further have the linear isomorphism @xmath113 , thus , as a result , given a full reference frame the screw gets represented by an element of @xmath114 in which the first three components are those of @xmath64 while the last three components are those of @xmath104 . given a screw @xmath115 , the _ screw axis _ of @xmath63 is the set of points for which the screw field has minimum module . the screw axis coincides with the set @xmath37 if @xmath116 and with a line of direction @xmath64 if @xmath72 . in both cases , if @xmath49 is any point in the screw axis then @xmath117 as a consequence , the screw axis is the set of points for which the screw field coincides with @xmath118 . for any point @xmath49 on the axis the motor reduction at @xmath49 is @xmath119 . let us observe that the former term in the right - hand side of eq . ( [ njf ] ) is proportional to @xmath64 and independent of the point , while the latter term is orthogonal to to @xmath64 and dependent on the point . let us suppose @xmath72 , the other case being trivial . let @xmath120 be any point , then it is easy to check that the axis which passes through @xmath49 in direction @xmath64 where @xmath121 is made of points @xmath122 for which @xmath123 . using the constitutive equation of screws we find that eq . ( [ njf ] ) holds . if @xmath67 is another point for which @xmath124 then that same equation gives @xmath125 , which implies that @xmath67 stays in the axis . thus the found axis is the locus of points @xmath67 for which @xmath124 . equation ( [ njf ] ) and the fact that @xmath126 imply that this axis is made of points for which the screw field is minimal . the other claims follow easily . usually the vector invariant and the screw axis are defined only for @xmath72 . however , we observe that it is convenient to extend the definition as done here in such a way that eq . ( [ njf ] ) holds for any screw . the case @xmath116 is admittedly special and can be called _ degenerate_. [ nhs ] the composition of applied vectors is nothing but the addition of the corresponding screws in the vector space @xmath73 . the resultant screw can then be represented with its motor in the screw axis which is given by the resultant @xmath64 aligned with the axis and the invariant vector @xmath118 having the same direction ( fig . [ fug ] ) . in this sense the composition of applied vectors does not give an applied vector . the operation of composition is closed only if the full space of screws is considered . ) .,width=340 ] two systems of applied vectors are said to be _ equivalent _ if they determine the same screw . one often looks for the simplest way of representing a screw by applied vectors . this is accomplished as follows . the screw is the sum of the free screw given by @xmath118 and the applied screw determined by the resultant @xmath64 applied on the screw axis . the free screw is generated by two opposite applied vectors @xmath127 , @xmath128 , placed in a plane perpendicular to @xmath118 and such that their magnitude times their arm gives @xmath129 . this reduces any screw to the sum of at most three applied screw ( two if @xmath116 ) . if @xmath72 the number can be reduced to two regarding @xmath64 as the sum @xmath130 , and absorbing one term of type @xmath131 through a redefinition of @xmath127 , and analogously for the other ( see fig . [ fug ] ) . the arm can be chosen in such a way that the resultants of the two applied screws are perpendicular . in summary any screw is generated by two applied screws whose resultants are either opposite with screw axes belonging to the same plane ( if @xmath116 ) , or equal in magnitude and perpendicular ( if @xmath72 ) . the _ pitch _ @xmath132 of a screw @xmath63 , with @xmath133 , is that constant such that @xmath134 . if @xmath116 and @xmath135 , we set by definition @xmath136 . clearly , for a non - trivial screw , the pitch vanishes if and only if the screw is an applied screw , and the pitch equals @xmath137 if and only if the screw is a free screw . the screws with a given pitch do not form a vector subspace . using the pitch the screw can be rewritten @xmath138,\ ] ] where @xmath139 , with @xmath140 normalized vector and @xmath141 . the quantity @xmath142 is called _ amplitude of the screw_. it must be said that for sir r. s. ball @xcite the screw is @xmath143 . however , it is not particularly convenient to regard @xmath143 as a fundamental object since these type of normalized screws do not form a vector space . sir r. s. ball would refer to our screws as _ screw motions_. we prefer to use our shorter terminology ( shared by @xcite ) because , for a dynamical screw @xmath31 , which we shall later introduce , no actual motion needs to take place . note also that the normalization of the pitch is chosen in such a way that , integrating the screw vector field by a parameter @xmath144 , i.e. by making a full rotation , one gets a diffeomorphism which is a translation by @xmath145 along the screw axis . in other words , with the chosen normalization , the pitch gives the translation of the screw for any full rotation . every screw is a vector field , thus we can form the lie bracket @xmath146 $ ] of two screws @xcite . in this section we check that this commutator is itself a screw and calculate its resultant . [ mjc ] the lie bracket @xmath147 $ ] is a screw with resultant @xmath148 and satisfies @xmath149 . some authors define the commutator of two screws as minus the lie bracket . let @xmath74 and @xmath75 be two screws @xmath150 let us fix a cartesian coordinate system @xmath151 , then the lie bracket reads @xmath152 note that @xmath153 $ ] which , using eq . ( [ kgu ] ) becomes @xmath154^i$ ] . inverting the roles of @xmath74 and @xmath75 we calculate the second term , thus we obtain the interesting expression @xmath155 let us check that it is a screw , indeed @xmath156-{\bm{s}}_1\times [ s_2(p)-s_2(q ) ] \\ & = { \bm{s}}_2\times[{\bm{s}}_1\times ( p - q)]-{\bm{s}}_1\times[{\bm{s}}_2\times ( p - q)]\\ & = [ { \bm{s}}_2\cdot(p - q))]{\bm{s}}_1-[{\bm{s}}_1\cdot(p - q))]{\bm{s}}_2=(-{\bm{s}}_1\times { \bm{s}}_2)\times ( p - q),\end{aligned}\ ] ] which proves also that the resultant is as claimed . the relation between the commutator and the scalar product is clarified by the following result [ mns ] let @xmath157 , be three screws , then @xmath158\rangle + \langle [ s_3,s_1 ] , s_2\rangle=0.\ ] ] furthermore , the quantity @xmath159\rangle $ ] reads @xmath158\rangle = s_3(p)\cdot ( { \bm{s}}_1\times { \bm{s}}_2)+s_2(p)\cdot({\bm{s}}_3\times { \bm{s}}_1)+s_1(p)\cdot({\bm{s}}_2\times { \bm{s}}_3),\ ] ] is independent of @xmath67 , and does not change under cyclic permutations of its terms . we use eq . ( [ jfp ] ) @xmath160\rangle & = { \bm{s}}_1\cdot[{\bm{s}}_2\times s_3(p)-{\bm{s}}_3\times s_2(p)]+s_1(p)\cdot(-{\bm{s}}_3\times { \bm{s}}_2 ) \\ & = s_3(p)\cdot ( { \bm{s}}_1\times { \bm{s}}_2)+s_2(p)\cdot({\bm{s}}_3\times { \bm{s}}_1)+s_1(p)\cdot({\bm{s}}_2\times { \bm{s}}_3).\end{aligned}\ ] ] this expression changes sign under exchange of @xmath74 and @xmath75 , thus we obtain the desired conclusion . given a screw @xmath115 it is possible to construct the linear map @xmath161 which is an element of the dual space @xmath162 . the linear map @xmath163 sends every screw to zero ( namely , it is the null map ) , if and only if @xmath164 . if @xmath63 is such that @xmath165 , then the scalar product with the free screw @xmath166 , shows that @xmath167 , a contradiction . if @xmath63 is a constant screw with vector invariant @xmath118 , then the screw scalar product with the applied screw @xmath168 , where @xmath49 is some point , gives @xmath169 , hence @xmath170 and thus @xmath63 is the null screw . we have shown that the linear map @xmath171 is injective . we wish to show that @xmath171 is surjective , namely any element of the dual vector space @xmath172 , can be regarded as the scalar product with some screw . we could deduce this fact using the injectivity and the equal finite dimensionality of @xmath73 and @xmath172 , but we shall proceed in a more detailed way which will allow us to introduce a useful basis for the space of screws and its dual . let us choose @xmath173 , and let @xmath55 be a positive oriented orthonormal base for @xmath40 , where @xmath41 denotes the orientation . namely , assume that we have made a choice of reference frame . the six screws , @xmath174 , @xmath175 , @xmath176 generate the whole space @xmath73 . indeed , if @xmath63 is a screw and @xmath177 is its motor at @xmath49 , @xmath178 , @xmath179 , then @xmath180 $ ] . as a consequence , every reference frame establishes a bijection between the screw space @xmath73 and @xmath114 as follows @xmath181 { \ reference \ frame \ } \begin{pmatrix } \bar{a } \\ \bar{b } \end{pmatrix } \in \mathbb{r}^6\ ] ] where @xmath182 ( vectors in @xmath183 are denoted with a bar , while the boldface notation is reserved for vectors in @xmath38 ) . the screw scalar product between @xmath184 in this representation takes the form @xmath185 thus the screw scalar product quadratic form is given by the @xmath186 matrix @xmath187 where @xmath188 is the identity @xmath189 matrix . let us now consider the six linear functionals @xmath190 , @xmath191 , @xmath176 . from the definition of scalar product evaluated at @xmath49 it is immediate that @xmath192 that is @xmath193 is the dual base to @xmath194 . every element @xmath195 is uniquely determined by the values @xmath196 , @xmath176 , that it takes on the six base screws @xmath197 , @xmath176 . by the above formulas , the linear combination @xmath198=\langle \sum_{i=1}^3 [ c_i m_i+d_i f_i ] , \cdot \rangle\ ] ] takes the same values on the screw base and thus coincides with @xmath199 . we can therefore establish a bijection of the dual space @xmath162 with @xmath114 as follows @xmath200 { \ reference \ frame \ } \begin{pmatrix } \bar{c } \\ \bar{d } \end{pmatrix } \in \mathbb{r}^6\ ] ] where @xmath201 ( it is convenient to distinguish this copy of @xmath114 with that isomorphic with @xmath73 introduced above ) . as a consequence the linear map @xmath202 , from @xmath73 to @xmath162 is bijective . thanks to this result any screw can be regarded either as an element of @xmath73 or , acting with the screw scalar product , as an element of @xmath162 . it must be stressed that if @xmath203 is represented by @xmath204 then @xmath205 is represented by @xmath206 , that is , the map from @xmath73 to @xmath172 which sends @xmath63 to @xmath163 is given in this representation by the matrix @xmath207 . the pairing between the elements of @xmath172 and those of @xmath73 is the usual one on @xmath114 . nevertheless , it is useful to keep in mind that we are actually in presence of two copies of @xmath114 ( as we consider two isomorphisms ) , the former isomorphic with @xmath73 and the latter isomorphic with @xmath162 . all this reduction to @xmath114 depends on the reference frame . as mentioned in the introduction most references of screw theory introduce the screw starting from its reduction or using a descriptive approach ( the screw has an axis , a pitch , etc . ) . as we argued in the introduction , it is pedagogically and logically preferable to define the screw without making reference to any reference frame . for future reference we calculate , using prop . [ mjc ] , the commutator between the screw base elements @xmath208&=0,\\ [ f_i , m_j]&=-[m_j , f_i]=-\sum_k \epsilon_{ijk } m_k , \\ [ f_i , f_j]&=-\sum_k \epsilon_{ijk } f_k.\end{aligned}\ ] ] the reader will recognize the lie algebra commutation relations of the group @xmath209 of rigid maps . we shall return to this non accidental fact later on . given a screw @xmath63 we consider the map @xmath210 which acts as @xmath211 $ ] . clearly , @xmath212 and if @xmath213 are screws , the jacobi identity for the lie bracket of vector fields @xmath214+[z,[x , y]]+[y,[z , x]]=0 $ ] , becomes @xmath215 let an origin @xmath11 be given and let us use the isomorphism with @xmath93 . let @xmath63 be represented by @xmath216 . if we introduce a full reference frame it is possible to check with a little algebra that , according to the above commutations , the map @xmath217 is represented by the matrix @xmath218 where for every @xmath219 , @xmath220 is an endomorphism of @xmath38 induced by the vector product . of course , if we had kept the reference frame @xmath114 isomorphism , then , as it is customary , with @xmath221 we would mean the @xmath222 matrix @xmath223 . a _ rigid motion _ is a continuous map @xmath224\times e \to e$ ] , which preserves the distances between points , i.e. for every @xmath61 , @xmath225 $ ] , we have @xmath226 , and such that @xmath227 is the identity map . a _ rigid map _ is the result of a rigid motion , that is a map of type @xmath228 . it can be shown that every rigid map is an affine map which preserves the scalar product and is orientation preserving ( * ? ? ? the rigid maps form a group usually denoted @xmath209 . in kinematics the velocity field of bodies performing a rigid motion satisfies the _ fundamental formula of the rigid body _ @xmath229 this formula is usually deduced from poisson formula for the time derivative of a normalized vector : @xmath230 . equation ( [ kih ] ) defines a screw which is called _ twist _ in the literature . let us denote this screw with @xmath231 , then @xmath232 and @xmath233 , where @xmath19 is the angular velocity of the rigid body . the instantaneous axis of rotation is by definition the screw axis of @xmath231 . let us recall that if a point moves with respect to a frame @xmath15 which is in motion with respect to a frame @xmath16 , then the velocity of the point with respect to @xmath16 is obtained by summing the drag velocity of the point , as if it were rigidly connected with frame @xmath15 , with the velocity relative to @xmath15 . if two kinematical screws are given and summed then the result gives a velocity field which represents ( by interpreting one of the screw field as the velocity field of the points at rest in @xmath15 with respect to @xmath16 ) the composition of two rigid motions . the nice fact is that the result is independent of which screw is regarded as describing the motion of @xmath15 . in other words the result has an interpretation in which the role of the screws can be interchanged . more generally , one may have a certain number of frames @xmath234 , @xmath235 , of which we know the screw @xmath236 which describes the rigid motion of @xmath237 with respect to @xmath234 . the motion of @xmath238 with respect to @xmath239 is then described by the screw @xmath240 . in particular , since the map which sends a screw to its resultant is linear , the angular velocity of @xmath238 with respect to @xmath239 is the sum of the angular velocities : @xmath241 . as illustrated in the introduction , the screw approach tells us something more . indeed , one can establish the direction of the instantaneous axis of rotation of @xmath238 with respect to @xmath239 by using the same methods used to determine the central axis in a problem of applied forces . indeed , we shall see in a moment that there is a parallelism between forces and angular velocities as they are both resultants of some screw . in dynamics the most important screw is that given by the moment field , and is called _ wrench_. let us recall that the momentum @xmath242 of a set of applied forces @xmath243 with respect to a point @xmath49 is given by @xmath244 if we consider @xmath67 in place of @xmath49 we get @xmath245 where @xmath246 is the force resultant . this equation shows that we are in presence of a screw @xmath31 such that @xmath247 , @xmath248 . the central axis of a system of forces is nothing but the screw axis . another example of screw is given by the angular momentum field . the angular momentum @xmath249 of a system of point particles located at @xmath250 with momentum @xmath251 with respect to a point @xmath49 is given by @xmath252 if we consider @xmath253 in place of @xmath49 we get @xmath254 where @xmath255 is the total linear momentum . this equation shows that we are in presence of a screw @xmath256 such that @xmath257 , @xmath258 . let us consider the constitutive equation of the screw of angular momentum @xmath259 the vector @xmath260 changes in time as the distribution of velocity and mass changes . actually , we can consider here another source of time change if we allow the point @xmath253 to change in time . let us first consider the case in which the angular momentum is considered with respect to a fixed point . by differentiating the previous equation with respect to time we get equation ( [ cgb ] ) . in other words the dynamic screw @xmath31 is the time derivative of the dynamic screw @xmath256 @xmath261 we use here a partial derivative to remind us that the poles are fixed . this equation replaces the first and second cardinal equation of mechanics . indeed , as the map @xmath262 is linear it follows @xmath263 which is the first cardinal equation @xmath264 in disguise . ( alternatively , write @xmath265 and differentiate ) . here the partial derivative coincides with the total derivative because the resultant is a free vector , it does not depend on the point . the second cardinal equation with respect to a point @xmath11 @xmath266 is obtained by evaluating eq . ( [ vgt ] ) at the point @xmath11 . in eq . ( [ vgt ] ) we have differentiated with respect to time assuming that the point with respect to which we evaluate the angular momentum does not change in time . in other words we have adopted a eulerian point of view . suppose now that on space we have a vector field @xmath267 which describes the motion of a continuum ( not necessarily a rigid body ) . in this case we have to distinguish the eulerian derivative with respect to time , which we have denoted @xmath268 , from the lagrangian or total derivative with respect to time @xmath269 . according to the latter , the second cardinal equation of mechanics reads @xmath270 where @xmath271 is the moving pole . let us differentiate @xmath272 with respect to time using the lagrangian description , that is , assuming that @xmath253 and @xmath49 move respectively with velocities @xmath273 , @xmath274 , and considering the angular momenta with respect to the moving points . we obtain @xmath275 using the second cardinal equation ( [ nxg ] ) we find that this is the constitutive equation of the momentum screw . nevertheless , the total derivative of the angular momentum is not a screw . the relation between the partial and total derivative is as follows @xmath276 where @xmath277= \lim_{\epsilon\to 0}\frac{1}{\epsilon}[{\bm{p}}\times { \bm{v}}(b)\epsilon ] = { \bm{p}}\times { \bm{v}}(b),\ ] ] thus @xmath278 however , suppose that the velocity field is itself a screw , @xmath279 so that the continuum moves rigidly , then from eq . ( [ bkx ] ) , using the previous results for commutators @xmath280-[\frac{{{\rm d}}{\bm{l}}(q)}{{{\rm d}}t}-{\bm{\omega}}\times { \bm{l}}(q)]\\ & = { \bm{f}}\times ( b - q)+[{\bm{p}}\times { \bm{v}}(b)-{\bm{\omega}}\times { \bm{l}}(b)]-[{\bm{p}}\times { \bm{v}}(q)-{\bm{\omega}}\times { \bm{l}}(q)]\\ & = { \bm{f}}\times ( b - q)+ [ k , l](b)-[k , l](q)={\bm{f}}\times ( b - q)+(-{\bm{\omega}}\times { \bm{p}})\times ( b - q)\\ & = [ { \bm{f}}-{\bm{\omega}}\times { \bm{p}}]\times ( b - q).\end{aligned}\ ] ] the time derivative @xmath281 with respect to the moving frame reads by poisson formula , @xmath282 , thus the previous result can be summarized as follows let us denote with @xmath269 the total derivative with respect to points that move rigidly according to a kinematical screw @xmath231 of vector field @xmath274 , and with @xmath281 the time derivative relative to the corresponding rigidly moving frame . the quantity @xmath283 defines a screw with resultant @xmath284 . this screw coincides with the screw @xmath285(q),\ ] ] thus @xmath286.\ ] ] we have only to prove the last statement , which follows easily from eq . ( [ kff ] ) and the definition of commutator . it must be remarked that in the previous result the angular momentum @xmath287 is calculated as in the original inertial frame , and not using the point particle velocities as given in the moving frame of twist @xmath231 . nevertheless , the previous result is quite interesting as it gives a dynamical application of the commutator . given a rigid body the kinematical screw @xmath231 fixes the velocity of every point of the rigid body and hence determines the angular momentum screw @xmath256 . the map @xmath288 is linear and is an extension of the momentum of inertia map which includes the translational inertia provided by the mass . let us recall that given some continuum and fixed a point @xmath49 , the _ momentum of inertia map _ @xmath289 , @xmath290 , is the linear map defined by the expression @xmath291\ ] ] where we have discretized the continuum into point masses @xmath292 located , respectively , at positions @xmath250 . let @xmath5 be the center of mass , namely the point defined by @xmath293 where @xmath294 . it is easy to prove the huygens - steiner formula @xmath295.\ ] ] let us consider a rigid motion described by a kinematical screw . from eq . ( [ bas ] ) @xmath296 where @xmath297 is the kinematical screw . it is clear that this map sends a screw into what has been proved to be another screw , and that this map depends on the location of the masses . let us differentiate with respect to time the equation @xmath298 and use @xmath299 to get @xmath300={\bm{v}}(o)+ { \bm{\omega}}\times ( c - o)={\bm{v}}(c).\ ] ] from this equation we also obtain @xmath301 . we have by writing @xmath302 @xmath303\\ & = ( m{\bm{v}}(c))\times ( q - c)+ i_c({\bm{\omega}}).\end{aligned}\ ] ] this equation shows how the kinematical screw determines the dynamical screw @xmath256 . plugging @xmath304 we find @xmath305 , and we reobtain , as we already know , @xmath306 . according to eq . ( [ nxc ] ) the screw axis of the angular momentum screw passes through the point @xmath307 with direction @xmath308 while the instantaneous axis of rotation passes through the point @xmath309 and has direction @xmath19 . let us consider a rigid body and let us decompose it into point particles of mass @xmath292 located at @xmath250 with mass velocity @xmath310 and momentum @xmath311 . let @xmath49 be any point @xmath312=\frac{1}{2 } \langle k , l \rangle.\end{aligned}\ ] ] thus the kinetic energy is one half the screw scalar product between the kinetic screw and the angular momentum screw . let us now suppose that on each point particle of mass @xmath292 located at @xmath250 acts a force @xmath313 possibly null . the power of the applied forces is the sum of the powers of the single forces . denoting with @xmath314 the work done by the forces @xmath315\\ & = { \bm{f}}\cdot { \bm{v}}(o)+ \sum_i { \bm{f}}_i\cdot [ { \bm{\omega}}\times ( r_i - o ) ] = { \bm{f}}\cdot { \bm{v}}(o)+ \sum_i { \bm{\omega } } \cdot [ ( r_i - o)\times { \bm{f}}_i]\\ & = { \bm{f}}\cdot { \bm{v}}(o)+ { \bm{\omega } } \cdot { \bm{m}}(o)=\langle k , d \rangle , \end{aligned}\ ] ] that is , the total power is the screw scalar product between the kinematical screw and the dynamical screw . since the scalar product is independent of @xmath11 , let us choose @xmath316 @xmath317 where we used knig s decomposition theorem . from the kinetic energy theorem we know that the variation of kinetic energy equals the work done by the forces on the rigid body , thus we expect that the last term vanishes . this is indeed so thanks to the following lemma . for a rigid body @xmath318 . let us fix a base for the vector space @xmath38 so that @xmath319 becomes represented by a time dependent matrix @xmath320 where @xmath321 is the diagonal ( time independent ) matrix of the principal moments of inertia and @xmath271 is a time dependent matrix giving the rotation of the principal directions of inertia with respect to a chosen fixed base of @xmath38 . differentiating with respect to time we obtain @xmath322 where @xmath323 is an antisymmetric matrix . however , @xmath19 belongs to the kernel of this matrix from which the desired result follows . the screw is particularly useful when modeling workless constraints between rigid bodies ( think for example of a robotic arm and at its constituent rigid parts ) . indeed , suppose that the body is made of @xmath324 rigid parts and let us focus on part @xmath325 . the constraints will reduce the possible motions of part @xmath325 for a given position of the other parts . in particular , the possible kinematical status of part @xmath325 for any given relative configuration of all parts will be described by a screw subspace @xmath326 of all the possible twists of part @xmath325 . let @xmath327 and let @xmath31 be the wrench acting on rigid body @xmath325 as a result of the interaction with the neighboring bodies . since , by assumption , the constraints are workless we must have @xmath328 . we conclude that the vector subspace @xmath329 made of the screws which are screw - orthogonal to the allowed movements ( i.e. screw - orthogonal to @xmath330 ) , is made by all the possible wrenches acting on body @xmath325 so as to make no work . two screws with vanishing screw scalar product ( i.e. screw - orthogonal ) are said to be reciprocal and the subspace @xmath331 is said to be reciprocal to subspace @xmath330 . let @xmath332 be a rigid map , namely the result of a rigid motion as it has been defined in section [ njc ] . we shall omit the proof that @xmath333 is an affine map such that @xmath334 where @xmath335 is a linear map which preserves the scalar product and the orientation . the rigid maps form a group denoted @xmath209 ( which sometimes we shall simply denote @xmath336 ) . in the coordinates induced by a reference frame @xmath337 has coordinates @xmath338 which are related to those of @xmath67 by @xmath339 where @xmath11 is a special orthogonal @xmath340 matrix . this expression clarifies that @xmath209 is a lie group . the three euler angles determining @xmath11 and the three translation coefficients @xmath341 , can provide a coordinate system on the 6-dimensional group manifold . we stress that we regard @xmath209 as an abstract lie group , and we do not make any privileged choice of coordinates on it ( we do not want to make considerations that depend on the choice of reference coordinates ) . the lie algebra @xmath342 ( which sometimes we shall simply denote @xmath343 ) of @xmath209 is the family of left invariant vector fields on @xmath209 . the lie commutator of two lie algebra elements is still an element of the lie algebra . this structure can be identified with the tangent space @xmath344 , @xmath345 being the identity element on @xmath336 , endowed with the lie bracket @xmath346:tg_e\times tg_e \to tg_e$ ] . let @xmath347 be the above ( left ) group action on @xmath37 so that @xmath348 . each point @xmath56 induces an _ orbit map _ @xmath349 given by @xmath350 , thus @xmath351 . we are interested on @xmath352 at @xmath353 , so that @xmath354 . if @xmath355 then @xmath356 , gives , for every @xmath56 , a vector field on @xmath37 which is the image of the lie algebra element @xmath357 . such vector fields on @xmath37 are called _ fundamental vector fields_. let us consider the exponential @xmath358 which is obtained by the integration of the vector field @xmath357 from @xmath345 . the orbit @xmath359 passing through @xmath67 is obtained from the integration of the vector field @xmath63 starting from @xmath67 . in other words , the 1-parameter group of rigid maps @xmath360 coincides with the 1-parameter group of diffeomorphisms ( which are rigid maps ) generated by the vector field @xmath63 . conversely , every such 1-parameter group of rigid maps determines a lie algebra element . since every screw element , once integrated , gives a non - trivial 1-parameter group of rigid maps , every screw is the fundamental vector field of some lie algebra element . that is , the map @xmath361 is surjective . but @xmath343 and @xmath73 are two vector spaces of the same dimensionality , thus this map is also injective . in summary , the screws are the representation on @xmath37 of the elements of @xmath342 . it is particularly convenient to study the lie algebra of @xmath209 through their representative vector fields on @xmath37 , indeed many features , such as the existence of a screw axis for each lie algebra element , become very clear . we can now use several results from the study of lie groups and their actions on manifolds @xcite . a central result is that the bijective map @xmath362 is linear and sends the lie bracket to the commutator of vector fields on @xmath37 . in some cases the exponential map from the lie algebra to the lie group is surjective ( which is not always true as the example of @xmath363 shows @xcite ) . this is the case of the group @xmath209 ( see ( * ? ? ? 2.9 ) ) thus , since every element of the lie algebra @xmath342 corresponds to a screw @xmath63 , and the exponential map corresponds to the rigid map obtained from the integration of the screw vector field by a parameter 1 , the surjectivity of the exponential map implies that every rigid map can be accomplished as the result of the integration along a screw , or , which is the same , by a suitable rotation along an axis combined with the translation along the same axis . this is the celebrated chasles theorem @xcite ) reformulated and reobtained in the lie algebra language . let us recall that if @xmath364 , then the expression @xmath365 $ ] defines a linear map @xmath366 called _ adjoint endomorphism_. the trace of the composition of two such endomorphisms defines a symmetric bilinear form @xmath367 called _ killing form _ on @xmath343 . the killing form is a special _ invariant bilinear form _ @xmath49 on @xmath343 , namely it is bilinear and satisfies ( use eq . ( [ bjz ] ) ) @xmath368 proposition [ mns ] shows that the scalar product of screws provides an invariant symmetric bilinear form on the lie algebra . we wish to establish if there is any connection with the killing form . as done in section [ gvo ] let us introduce an origin @xmath11 and use the isomorphism of @xmath73 with @xmath93 . if @xmath369 are represented by @xmath370 and @xmath371 , then the screw scalar product is @xmath372 according to the result of section [ gvo ] we have @xmath373 we conclude that the killing form is an invariant bilinear form which is distinct from the screw scalar product . it coincides with the killing form of the lie group of rotations alone and thus , it does not involve the translational information inside the @xmath11-terms . therefore , the screw scalar product provides a new interesting invariant bilinear form , which is sometimes referred to as the _ klein form _ of @xmath342 . from eq . ( [ lmu ] ) we find that @xmath374 is represented by @xmath375 from which , using the symmetry properties of the mixed product , we can check again that the screw scalar product is invariant @xmath376\\ & + [ ( { \bm{y } } \times { \bm{z}})\cdot { \bm{x}}^{o}+({\bm{y}}^{o}\times { \bm{z}})\cdot { \bm{x } } + ( { \bm{y } } \times { \bm{z}}^{o})\cdot { \bm{x}}]=0.\end{aligned}\ ] ] screw theory , although venerable , has found some difficulties in affirming itself in the curricula of the physicist and the mechanical engineer . this has changed in the last decades , when screw theory has finally found application in robotics , where its ability to deal with the composition of rigid motions has proved to be much superior with respect to treatments based on euler coordinates . we have given here a short introduction to screw theory which can provide a good starting point to a full self study of the subject . we started from a coordinate independent definition of screw and we went to introduce the concepts of screw axis , screw scalar product and screw commutator . we introduced the dual space and showed that any frame induces an isomorphism on @xmath114 which might be used to perform calculations . we then went to consider kinematical and dynamical examples of screws , reformulating the cardinal equations of mechanics in this language . particularly important was the application of the screw scalar product in the expressions for the kinetic energy and power , in fact the virtual work ( power ) is crucial in the formulation of lagrangian mechanics . in this connection , we mentioned the importance of reciprocal screws . finally , we showed that the space of screws is nothing but the lie algebra @xmath342 , and that the screw scalar product is the klein form . philosophically speaking , screw theory clarifies that the most natural basic dynamical action is not the force , but rather the force aligned with a mechanical momenta ( remark [ nhs ] ) . in teaching we might illustrate the former action with a pushing finger and the latter action with a kind of pushing hand . analogously , the basic kinematical action is not given by the act of pure rotation ( or translation ) but by that of rotation aligned with translation . again , for illustration purposes this type of motion can be represented with that of a ( real ) screw . clearly , in our introduction we had to omit some arguments . for instance , we did not present neither the cylindroid nor the calculus of screws . nevertheless , the arguments that we touched were covered in full generality , emphasizing the geometrical foundations of screw theory and its connection with the lie group of rigid maps . we hope that this work will promote screw theory providing an easily accessible presentation to its key ideas . i thank d. zlatanov and and m. zoppi for suggesting some of the cited references . this work has been partially supported by gnfm of indam .

this work introduces screw theory , a venerable but yet little known theory aimed at describing rigid body dynamics . this formulation of mechanics unifies in the concept of screw the translational and rotational degrees of freedom of the body . it captures a remarkable mathematical analogy between mechanical momenta and linear velocities , and between forces and angular velocities . for instance , it clarifies that angular velocities should be treated as applied vectors and that , under the composition of motions , they sum with the same rules of applied forces . this work provides a short and rigorous introduction to screw theory intended to an undergraduate and general readership . keywords : rigid body , screw theory , rotation axis , central axis , twist , wrench . + msc : 70e55 , 70e60 , 70e99 .

as mentioned in the main text , an important property of the tljs is that their ends localize exotic forms of zero modes , called parafermion zero modes , which generalize the usual majorana zero mode algebra . @xcite here we will briefly recall the definition and physical meaning of these zero mode operators , reviewing results from @xcite . let us imagine aligning all tljs , and cutting the whole system along the line , yielding the counterpropagating edge states . then , the normal regions of the fluid can be obtained by gluing the edges back together with the electron tunneling terms @xmath134 , while the twisted line junctions can be described by gluing the edges back together with the twisted tunneling @xmath135 , where we have defined @xmath136 , and @xmath137 . here , we have included also the tunneling term @xmath138 , which is not present in the experimental setup of fig . [ expsetup ] . the ends of the tlj can therefore be considered as domain walls between these two different ways of gapping the counterpropagating edge states : @xmath139 where @xmath140 indicates the normal regions , while @xmath141 are the regions associated with the tljs . as shown in @xcite , we can define quasiparticle tunneling operators ( see fig . [ edgefig ] ) @xmath142 , @xmath143 , which are zero modes of the edge hamiltonian , @xmath144 = 0 $ ] , and which satisfy a @xmath145 parafermion algebra : @xmath146 , and @xmath147 . the defects ( genons ) can be viewed as domain walls between two different ways of generating an energy gap in the counterpropagating edge states ( double arrows indicate interedge electron tunneling ) . the zero mode operator @xmath148 corresponds to the quasiparticle hopping process shown in the figure , projected into the ground state subspace of the edge theory . , width=288 ] other ways of creating domain walls between two different ways of generating an energy gap in a single pair of counterpropagating chiral luttinger liquids has been explored in @xcite , where similar parafermion zero modes are obtained . in the experimental setup described in the main text , there are gapless edge states surrounding the tljs . these edge states do not affect the topological degeneracies , the exponential localization of the zero modes , or the possibility of braiding the genons by tuning the interactions between the parafermion zero modes . nevertheless , it might be desirable to remove these gapless edge states , while still preserving the topological degeneracy . this can be done simply by alternating the offset in space ( see fig . [ zigzag ] ) . the tunneling term in the edge states can be modelled as follows : @xmath149 while @xmath150 and @xmath151 are never both non - zero simultaneously , at long wavelengths only the average value of @xmath152 and @xmath153 is important . therefore , both edge states can be gapped in the twisted way on average in this setup . alternating the offset of the top and bottom gate will allow us to generate an energy gap for all edge states . at long wavelengths , both twisted tunneling terms , @xmath154 are simultaneously present . , width=384 ] here , will present the derivation of the current noise cross - correlation for the interferometer summarized in the main text . we consider a tunneling hamiltonian with four qpcs @xcite : @xmath155 , \nonumber \\ h_i(t ) & = v_i \mathcal{o}^+(x_i ) + v_i^ * \mathcal{o}^-(x_i ) , \nonumber \\ \tilde{h}_i(t ) & = \tilde{v}_i \tilde{\mathcal{o}}^+(\tilde{x}_i ) + \tilde{v}_i^ * \tilde{\mathcal{o}}^-(\tilde{x}_i),\end{aligned}\ ] ] where @xmath156 @xmath157 are the tunneling matrix elements for the different qpcs , and @xmath158 , where @xmath159 are the potential differences across the qpcs and @xmath160 is the elementary fractional charge of the quasiparticles . we note that in the setup of fig . [ interferometer ] , to lowest order in the tunneling matrix elements , the voltage across each qpc is @xmath161 . for what follows , it will be useful to define : @xmath162 the operator @xmath163 is the quasiparticle tunneling operator across qpcs enclosing the normal regions of the fluid : @xmath164 where @xmath165 are klein factors that ensure that the quasiparticle tunneling operators at different points in space commute with each other , as required . the operators @xmath166 are quasiparticle tunneling operators across the twisted line junctions ( tljs ) : @xmath167 where @xmath168 are also klein factors . the commutation relations of @xmath169 and @xmath168 can be determined from the boson commutation relations : @xmath170 & = i \frac{\pi}{m } sgn(x - x ' ) \nonumber \\ [ \phi_{ri}(x),\phi_{ri}(x ' ) ] & = -i \frac{\pi}{m } sgn(x - x'),\end{aligned}\ ] ] from which we determine that @xmath171 = 0 $ ] , @xmath172 = 0 $ ] , and @xmath173 where @xmath174 . in the interferometer discussed in the main text , which probes the non - abelian character of the genons , we require @xmath175 . however it will also be helpful to compare this with the case @xmath176 , which does not probe the non - commutativity of the two non - commuting , non - contractible quasiparticle paths . now , we let @xmath177 and @xmath178 be the current that is backscattered through the normal regions and the tljs , respectively : @xmath179,\end{aligned}\ ] ] where @xmath180 is the total charge on the right edge in the @xmath181th layer . thus : @xmath182 = \frac{i}{m } \sum_{i=1}^2 [ v_i \mathcal{o}^+(x_i ) - v_i^ * \mathcal{o}^-(x_i ) ] = \frac{i}{m } \sum_{i=1}^2 \sum_{\epsilon_i = \pm } \epsilon_i v_i^{\epsilon_i } \mathcal{o}^{\epsilon_i}(x_i ) \nonumber \\ i_{b2}(t ) & = \frac{i}{m}\sum_{i=1}^2 [ \tilde{\gamma}_i e^{-i e^ * v_2 t } \tilde{\mathcal{o}}^+(\tilde{x}_i ) - h.c . ] = \frac{i}{m}\sum_{i=1}^2 [ \tilde{v}_i \tilde{\mathcal{o}}^+(\tilde{x}_i ) - \tilde{v}_i^ * \tilde{\mathcal{o}}^-(\tilde{x}_i ) ] = \frac{i}{m } \sum_{i=1}^2 \sum_{\epsilon_i = \pm } \epsilon_i \tilde{v}_i^{\epsilon_i } \tilde{\mathcal{o}}^{\epsilon_i}(\tilde{x}_i)\end{aligned}\ ] ] in the setup of fig . [ interferometer ] , we find that : @xmath183 where @xmath184 and @xmath185 are the currents flowing through the device in the first and second layer ( see fig . [ interferometer ] ) . in order to understand @xmath186 , observe that current can be backscattered through the tljs in two different ways . the first way is through quasiparticle tunneling across the tlj . the second way is if the outer edge that follows the constriction of the qpc overlaps the gapless edge states that flow around the tlj . in this case , there can be interlayer electron tunneling between the top and bottom outer edges of the sample , through the edge states surrounding the tlj . this latter contribution is what we call @xmath186 . therefore : @xmath187 we will be interested in the limit @xmath126 , in which case it will become clear that @xmath186 , which involves only electron tunneling , will not have long - time cross - correlations between the quasiparticle tunneling current @xmath177 . therefore , we will compute the following interlayer noise cross - correlation : @xmath188 note that the expectation values are given by @xmath189 where the @xmath190 indicates an expectation value in the unperturbed theory . @xmath191 indicates the keldysh contour , and @xmath192 is time - ordering in the keldsyh contour . in our calculation , we will drop the subscript and assume all expectation values are in the unperturbed theory . let us first present the result of the calculation . we find that to lowest oder in the qpc tunneling matrix elements , @xmath193.\end{aligned}\ ] ] where we have defined the dimensionless quantities @xmath194 and @xmath195 and @xmath196 are the phase differences between the tunneling matrix elements of the qpcs : @xmath197 , and @xmath198 . note that @xmath128 depend on only dimensionless quantities . @xmath199 , \nonumber \\ s_{12}^{(2)}(\bar{\omega}_1 , \bar{\omega}_2 , x , y , \tilde{\varphi}_{12 } , \varphi_{12 } ) = & 8 c_r(\bar{\omega}_1 , 2 x ) \cos \tilde{\varphi}_{12 } [ 2 ( \cos(\pi / m ) - 1 ) \cos(\varphi_{12 } ) c_r(\bar{\omega}_2,2 y ) \nonumber \\ & - \sin(\pi / m ) \sin(\varphi_{12 } ) f_+(\bar{\omega}_2 , y)],\end{aligned}\ ] ] the functions @xmath200 are defined as @xmath201 \left ( \frac{1}{\sinh |t + \bar{x}| \sinh |t - \bar{x}| } \right)^{1/m}\end{aligned}\ ] ] while @xmath202 is defined as @xmath203 \sinh [ t+ |b| ] } \right)^{1/m } \right\}\end{aligned}\ ] ] as a check , observe that when @xmath204 and @xmath205 , the two non - contractible paths for quasiparticle propagation commute , and therefore this noise cross - correlation should vanish at long times , which it indeed does . in fig . [ plotsf ] , we plot the functions @xmath206 and @xmath207 . in fig [ plotss ] , we plot @xmath208 , for @xmath209 ; that is , for the case where the voltage differences between all qpcs is the same , @xmath210 , and for zero phase differences @xmath195 and @xmath196 . we see that when @xmath47 and @xmath48 are of order one , @xmath128 reach their maximum value , and decay quickly as @xmath47 and @xmath48 increase . from this , we see that the noise cross - correlation will require the dimensionless parameters @xmath211 , @xmath212 , and @xmath213 to all be of order one . if we assume @xmath214 , and @xmath215 , we see that this requires @xmath216 and @xmath217 be on the order of several @xmath218 , which is reasonable for realistic device dimensions . here , we will determine the keldysh - contour ordered correlation functions of the klein factors , which are required for the calculation . to do this , first we observe that the two point function for klein factors that commute with each other should be one . next , we observe that the correlation functions of several klein factors should factor into two - point functions . in order to obtain the two - point function of non - commuting klein factors , it is helpful to represent the klein factors in terms of bosonic operators @xcite : where @xmath220 are bosonic operators and for what follows we will leave the normal ordering @xmath221 implicit . using the fact that @xmath222}$ ] for cases where @xmath223 $ ] is a @xmath224-number , we find that the commutation relation of the klein factors is reproduced when @xmath225 = -i\pi sgn(x_i - \tilde{x}_j ) , \;\;\;\;\ ; [ \theta_i , \theta_j ] = [ \tilde{\theta_i } , \tilde{\theta}_j ] = 0\end{aligned}\ ] ] since @xmath226 and @xmath227 are conjugates , we can form a harmonic oscillator algebra : @xmath228 such that @xmath229 = - sgn(x_i - \tilde{x}_j ) , \;\;\ ; \langle a_{ij } a_{ij}^\dagger \rangle = \frac{1 - sgn(x_i - \tilde{x}_j)}{2 } .\end{aligned}\ ] ] we suppose that the hamiltonian for these bosonic fields is zero , as they have no dynamics . from this we conclude : @xmath230 from this , we conclude that the keldysh contour ordered correlation function is : @xmath231 where @xmath232 therefore : @xmath233 for later reference , we will need the following correlation function : @xmath234 to calculate this , we use the fact that @xmath235 , where in the klein factor case , we only include correlation functions between non - commuting variables . therefore : @xmath236 } & t > 0 \\ e^ { i \frac{\epsilon_1 \epsilon_2}{2 m } \pi [ \eta ( s_{\alpha_1 \beta_2 } - s_{\alpha_1 \alpha_2 } ) + \eta_1 ( s_{\beta_1 \alpha_2 } - s_{\beta_1 \beta_2 } ) ] } & t < 0 \\ \end{array}\right.\end{aligned}\ ] ] where @xmath237 . when @xmath176 , @xmath238 for all @xmath152 , @xmath153 , and therefore @xmath239 when @xmath175 , which is the case of interest for our proposed interferometer , we instead have @xmath240 } & \text { if } & t > 0 \\ e^{-i \frac{\epsilon_1 \epsilon_2}{m } ( \alpha_2 - \beta_2)\pi [ -\eta \delta_{\alpha_1 2 } + \eta_1 \delta_{\beta_1 2 } ] } & \text { if } & t < 0 \\ \end{array } \right.\end{aligned}\ ] ] we will need the following correlation functions : @xmath241 at finite temperatures , this is ( setting the velocities of the left- and right- movers to unity ) : @xmath242 \sin [ \pi t ( \delta + i \chi_{\eta_1 \eta_2}(t ) ( t - x ) ) ] } \right)^{1/m},\end{aligned}\ ] ] where @xmath243 taking the limit @xmath244 , we can replace the above by @xmath245 \sinh [ \pi t|t - x| ] } \right)^{1/m}\end{aligned}\ ] ] the chiral correlator , involving only the left or right - moving modes ( and re - inserting the velocities @xmath246 for the left / right movers ) , is : @xmath247 } \right)^{1/m}\end{aligned}\ ] ] here we describe in detail the calculation of @xmath248 , to lowest order in the tunneling matrix elements . we have ( recall eq . ( [ htunn ] ) , ( [ corrfun ] ) ): @xmath249 consider the correlation function : @xmath250 the above correlation function vanishes unless @xmath251 and @xmath252 . therefore : @xmath253 recall that in our notation , the superscript @xmath254 implies complex conjugation , not inverse . @xmath255 this gives : @xmath256 the green s functions @xmath257 decay exponentially in @xmath258 ; therefore , the integral will only be dominated by regions where @xmath259 and @xmath260 . furthermore , we are interested in the long time behavior , where @xmath258 is large . in this limit , the ratio in parantheses above will go to one : @xmath261 when @xmath262 and @xmath263 , @xmath264 . therefore : @xmath265 therefore , so far we have , for large @xmath258 , @xmath266 \nonumber \\ & \eta_1 \eta_2 \int_{-\infty}^{\infty } e^ { i \epsilon_2 \omega_2 ( -\tau_2 ) } g^{v_{2r};v_{1l}}_{-\eta \eta_2}(\tilde{x}_{\alpha_2 } - \tilde{x}_{\beta_2 } , -\tau_2 ) d\tau_2 \int_{-\infty}^{\infty } d\tau_1 e^{i \epsilon_1 \omega_1 ( - \tau_1 ) } g^{v_{1r};v_{1l}}_{\eta \eta_1 } ( x_{\alpha_1 } - x_{\beta_1 } , t - \tau_1)\end{aligned}\ ] ] in the case where @xmath176 , using the formula for the klein factor expression found earlier , we find that the difference of klein factor correlation functions vanishes , so that @xmath267 for @xmath126 . now let us focus on the case @xmath268 . in this case , for @xmath269 , @xmath270 } - 1 ] \nonumber \\ & \eta_1 \eta_2 \int_{-\infty}^{\infty } e^ { i \epsilon_2 \omega_2 ( -\tau_2 ) } g^{v_{2r};v_{1l}}_{-\eta \eta_2}(\tilde{x}_{\alpha_2 } - \tilde{x}_{\beta_2 } , -\tau_2 ) d\tau_2 \int_{-\infty}^{\infty } d\tau_1 e^{i \epsilon_1 \omega_1 ( - \tau_1 ) } g^{v_{1r};v_{1l}}_{\eta \eta_1 } ( x_{\alpha_1 } - x_{\beta_1 } , t - \tau_1)\end{aligned}\ ] ] then we can write the noise as a simple expression : @xmath274 } - 1 ] a_{\eta \eta_1}^{\epsilon_1 \omega_1}(x_{\alpha_1 \beta_1 } ) a_{-\eta \eta_2}^{\epsilon_2 \omega_2}(\tilde{x}_{\alpha_2 \beta_2})\end{aligned}\ ] ] and @xmath275 } - 1 ] a_{\eta \eta_1}^{\epsilon_1 \omega_1}(x_{\alpha_1 \beta_1 } ) a_{-\eta \eta_2}^{\epsilon_2 \omega_2}(\tilde{x}_{\alpha_2 \beta_2})\end{aligned}\ ] ] now let us try to make more explicit the above expressions . we start with the @xmath269 case . summing over @xmath276 and @xmath277 gives : @xmath278 for @xmath279 , we perform the sum over @xmath280 , @xmath281 to get : @xmath282 } - 1 ] a_{\eta \eta_1}^{\epsilon_1 \omega_1}(x_{\alpha_1 \beta_1 } ) ( a_{-\eta \eta_2}^{\epsilon_2 \omega_2}(\tilde{x}_{12 } ) - a_{-\eta \eta_2}^{-\epsilon_2 \omega_2}(\tilde{x}_{12}))\end{aligned}\ ] ] this can be written as the sum of two terms : @xmath283 where @xmath284 contains the terms where @xmath285 in the above sum . @xmath286 contains terms in the sum ( [ s12tpos ] ) where @xmath287 . similarly , we write @xmath288 where @xmath289 contains terms in the sum ( [ s12tneg ] ) where @xmath290 , while @xmath291 contains terms in the sum where @xmath292 . after simplifying , we find @xmath293 where @xmath128 are presented in ( [ corrresult ] ) .

the idea of topological quantum computation is to build powerful and robust quantum computers with certain macroscopic quantum states of matter called topologically ordered states . these systems have degenerate ground states that can be used as robust topological qubits " to store and process quantum information . however , a topological qubit has not been realized since the proposed systems either require sophisticated topologically ordered states that are not available yet , or require complicated geometries that are too difficult to realize . in this paper , we propose a new experimental setup which can realize topological qubits in a simple bilayer fractional quantum hall ( fqh ) system with proper electric gate configurations . compared to previous works , our proposal is accessible with current experimental techniques and only involves well - established topological states . our system can realize a large class of topological qubits , generalizing the majorana zero modes studied in the recent literature to more computationally powerful possibilities . we propose three tunneling and interferometry experiments to detect the existence and non - local topological properties of the topological qubits . one of the greatest challenges of modern physics is to harness the power of quantum mechanics in order to realize quantum computation . the fundamental obstacle is the difficulty of maintaining quantum coherence for long times . among the various approaches , topological quantum computation ( tqc)@xcite is a proposal which relates the protection of quantum coherence with fundamental properties of some macroscopic quantum states of matter , called topologically ordered states@xcite . these are gapped states of matter with a certain ground state degeneracy determined solely by the spatial topology of the system . all degenerate ground states are indistinguishable in all local properties , so that the quantum coherence between them is robust to environmental noise , for a time that scales exponentially in system size . in two - dimensions , all low energy excitations of topological states are point - like quasi - particles ; for a subset of them , called non - abelian states , a set of quasi - particles at fixed positions can also carry a robust , topological degeneracy of states . the topological degeneracy in both ground states and non - abelian quasi - particle states can be used as a topological qubit , " to carry quantum information . most proposals use the latter since it is considered easier to realize quantum operations by manipulating the quasiparticle positions . while the non - abelian states are not yet experimentally established , currently there are two types of candidate systems for non - abelian states : the non - abelian quantum hall states@xcite , the most promising of which is proposed to be realized at filling @xmath0 @xcite , and topological superconductors with majorana zero modes@xcite . the topological superconductor proposals have the advantage of possibly larger energy gaps and well - controlled location of non - abelian defects . however , the non - abelian defects that can be realized in these proposals are restricted to ising anyons , which have limited computational power @xcite , and the required proximity between superconductor and semiconductor or topological insulator adds complications to probing the non - abelian anyons . in this paper , we propose a new experimental platform that realizes a wide class of non - abelian defects @xcite , vastly generalizing majorana zero modes , using conventional abelian bilayer quantum hall states . our proposal is inspired by recent progress in vertical field effect transistors,@xcite and recent theoretical developments regarding extrinsic defects in fqh states.@xcite as is illustrated in fig . [ expsetup ] , our setup only requires the simplest bilayer fqh states that have been observed in the laboratory @xcite , with certain configurations of top and bottom gates . the key idea is to use the gate configuration to induce inter - layer tunneling and thus build staircases " that coherently connect the two layers . the end points of such staircase " lines are non - abelian defects , which we refer to as _ genons _ , although the bilayer fqh state itself is abelian . the non - abelian statistics of genons are determined by the parent state of the bilayer , which include the ising anyons realized in topological superconductors as a special case . with a larger gap than known non - abelian fqh states and well - controlled position of non - abelian defects , but without requiring superconductivity , our proposal combines the advantage of both types of proposed systems in the literature . in what follows , we explain the proposed experimental setup and discuss two novel experiments that can be used to detect the topological qubit . * experimental setup * our experimental setup ( fig . [ expsetup ] ) consists of a double - layer quantum well system , with line junctions in each layer that are offset in the lateral direction . one way to create the line junctions is with top and bottom gates ; the gates are used to deplete the layer closest to them , and to increase the interlayer tunneling over a small region in their vicinity . another way to create the line junctions , which has been employed successfully in single - layer systems,@xcite is through a physical obstruction within each layer , such as an undoped al@xmath1ga@xmath2as / asas barrier . when the layers are well - separated , each layer can individually form its own single - layer fqh state , such as a filling fraction @xmath3 laughlin fqh state . with larger interlayer interactions , more exotic interlayer correlated states are also possible . here we consider the simplest case where each layer individually forms a @xmath4 laughlin fqh state . double layer 1/3-laughlin states have routinely been realized experimentally , with gaps on the order of several k @xcite . at the boundary of the line junctions , there are gapless chiral edge states . with the offset between top and bottom gates , and in a suitable parameter region that will be analyzed below , it is possible to induce backscattering between counter - propagating edge states from _ different layers . the back - scattering can open a gap at the boundary and effectively build a staircase " that coherently connects the two layers ( fig . [ expsetup ] ( b ) ) . such staircases are the essential elements of our proposal , which enable the realization of non - planar geometries . as will be discussed below , the end points of a staircase line junction are non - abelian topological defects , a pair of which carries topological degeneracy and thus acts as a topological qubit . in the following we will refer to such a staircase as a twisted line junction ( tlj ) . _ * conditions for building tlj * the hamiltonian of the edge states around the line junctions is described by the chiral luttinger liquid theory @xcite , @xmath5 , with @xmath6 @xmath7 and @xmath8 are the hamiltonian densities of the kinetic / interaction terms and the backscattering terms , respectively . @xmath9 and @xmath10 denote the chirality and the layer index of the edge states , respectively . the densities @xmath11 , with @xmath12 the chiral boson fields satisfying the commutation relations @xmath13 = \pm i \frac{\pi}{m } \delta_{ij } sgn(x - y)$ ] . @xmath14 are back - scattering amplitudes between counter - propagating edge states @xmath15 and @xmath16 of the two layers ( see fig . [ expsetup ] ) . the velocity of the edge modes , @xmath17 , is assumed for simplicity to be the same for all of the edge states . the diagonal entries of @xmath18 , @xmath19 are normalized to @xmath20 , while the off - diagonal elements of the symmetric matrix @xmath18 parameterize the density - density interaction between the different edge channels . to realize the tlj , the following conditions are required : \1 ) _ the line junction of one layer must not destroy the fqh state directly beneath ( or above ) it . _ in the setup of fig . [ expsetup ] , this means the gates must deplete the mobile electron density in the layer closest to them , while keeping the density in the further layer such that it remains in its fqh plateau . based on previous experimental studies , such a situation is easily possible@xcite . \2 ) _ @xmath21 , _ i.e. _ the dominate back - scattering term should be @xmath22 . _ as long as the width of the line junction @xmath23 is significantly larger than the inter - layer distance @xmath24 , this condition can be easily satisfied . if the system is clean , the intra - layer tunneling has a phase factor @xmath25 ( with @xmath26 the magnetic length ) due to the momentum mismatch , while the vertical tunneling has no momentum mismatch . this further suppresses other back - scattering terms . furthermore , near the gates where the density is not depleted , the wave functions will extend further near the other layer , because the fringe fields of the gate raise the potential well that vertically confines the 2deg . this exponentially increases @xmath22 , as has been observed experimentally@xcite . \3 ) _ the back - scattering @xmath22 should generate an energy gap . _ in the absence of interactions between edge states , the tunnelings @xmath27 are irrelevant perturbations in the renormalization group ( rg ) sense . in this case , a large tunneling @xmath28 is required to open an energy gap and lead to the required tlj . however , it is possible for the screened coulomb interaction between edge states to make @xmath22 a relevant perturbation in the rg sense , so that any small @xmath22 will open a gap and allow coherent quasi - particle tunneling across the junction . the coulomb interaction gives an estimate of @xmath29 : @xmath30 , where @xmath31 is the coulomb energy between edges . when the gates are wide @xmath32 , the only appreciable interaction will be between the two edges on top of each other , which , ignoring screening due to the gates , is estimated by @xmath33 . in this case , the scaling dimension of the electron back - scattering term is @xmath34 @xcite . the tunneling is relevant in the rg sense if @xmath35 in the clean limit , or @xmath36 for disorder induced backscattering ( assuming gaussian - correlated disorder ) . for realistic parameters in gaas samples , @xmath37 , @xmath38 , @xmath39 , @xmath40 , we see that the repulsive interaction can easily make @xmath41 to be order one , so that the tunneling will be relevant . \4 ) _ the edges of the line junctions should be sharp enough in order for the edge theory description ( [ h0 ] ) , ( [ ht ] ) to be applicable . _ the width of the edge region @xmath42 ( fig . [ expsetup ] ( a ) ) , in which the density of electrons gradually decreases to zero , must be on the order of @xmath43 , so that it can be treated as being part of the gapless edge of the bulk fqh state . @xmath42 is set by the fringe fields of the gates , and is on the order of the vertical distance @xmath44 between the 2deg and the gate , so it is preferable to minimize @xmath44 . for this reason , in addition to gaas double quantum wells , two layers of graphene , separated by a thin insulating barrier may be an advantageous system , since @xmath44 can be made dramatically smaller @xcite . now that we have demonstrated the feasibility of building the tlj , we will discuss the realization of topological qubits and experimental predictions . * topological degeneracy : geometrical picture * a @xmath4 laughlin fqh state on a genus @xmath45 surface has a ground state degeneracy @xmath46 @xcite . this can be understood on a torus to be a consequence of the two independent non - contractible loops , @xmath47 and @xmath48 , which intersect only once , combined with the fractional statistics of the quasiparticles . together , these imply the following magnetic algebra for quasiparticle loop operators @xmath49 and @xmath50 : @xmath51 . @xmath52 measures the fractional charge inside the loop @xmath53 respectively , which can be shown to commute with the low energy hamiltonian in the ground state subspace . therefore , the ground states form an irreducible representation of this algebra , which is at least @xmath54-dimensional . a genus @xmath45 surface gives @xmath45 independent copies of the magnetic algebra , and therefore a topological degeneracy of @xmath46 . on a finite size system , quasiparticle - quasihole tunneling around @xmath47 , @xmath48 mixes the different states , splitting the degeneracy by an energy that is at most exponentially small in system size . a pair of tljs introduces two non - contractible loops , so the space is topologically equivalent to a torus . the edge states around the outer edge of the sample and around the gates are mapped to edge states around four holes on the surface of the torus . , width=316 ] now consider our experimental proposal ( fig . [ expsetup ] ) . with two tlj s ( fig . [ doublestaircase ] ) , the system now has two independent non - contractible loops @xmath47 and @xmath48 that intersect only once ; the system is topologically equivalent to a torus with four holes cut out on the surface ( fig . [ doublestaircase ] ) . with @xmath55 tlj s , the system is topologically equivalent to a genus @xmath56 surface , with @xmath57 holes cut out on the surface , which will have a topological degeneracy of @xmath46 . the holes cut on the surface contribute chiral edge states with a finite size gap inversely proportional to the perimeter of the hole , but they do not affect the topological degeneracies at all . we note that the gapless edge states surrounding the tljs can also be removed through more complicated interdigitated gate configurations ( see supplemental materials ) . since each tlj increases the degeneracy by a factor of @xmath54 , each end point of the tlj , viewed as a point defect , can be associated with @xmath58 degrees of freedom . similar to the majorana zero mode carrying @xmath59 degrees of freedom , each end point of a tlj is a non - abelian topological defect which we refer to as a _ genon_.@xcite the resulting topological degeneracies can be used to realize a topological qubit ; this setup enables manipulations of such topological qubits by braiding the genons . the braiding process can be shown to correspond to topologically non - trivial surgery operations of the genus @xmath45 surface . @xcite these mathematical operations correspond to cutting the space along a non - contractible loop , rotating by @xmath60 , and gluing back together . * proposed experiments * _ 1 . quantized zero - bias quasiparticle conductance _ the @xmath58 topological degeneracy associated with a genon , _ i.e. _ an end point of the tlj , corresponds to a localized parafermion zero mode@xcite , which generalizes the usual majorana zero modes in topological superconductors . zero mode operators @xmath61 can be defined at each genon , which satisfy the parafermion algebra @xmath62 with a proper ordering of all genons in the system . physically , the operator @xmath61 is a quasi - particle tunneling operator between the two layers at the genon point , which commutes with the hamiltonian . we refer to ref . @xcite for more rigorous definition of the zero mode operators . in the case of the majorana zero modes localized at the ends of a 1d topological superconductor , one of its key signatures is a quantized zero bias conductance peak with conductance @xmath63 when a normal metal lead is coupled to the majorana zero mode @xcite . this is due to the perfect andreev reflection in the presence of the majorana mode . in our system , the analog of an electron in topological superconductor is a quasi - particle - quasi - hole pair , _ i.e. _ , a topological exciton " with charge @xmath64 . here the two numbers represent the charge of the exciton in the two layers . when such an exciton moves across the tlj , the two layers are exchanged , such that its charge becomes @xmath65 ; a topological exciton will become its anti - particle when going across a tlj , just like an electron being transformed into a hole in a superconductor through an andreev process . the topological exciton has a perfect andreev " reflection at presence of the parafermion zero mode , which can be probed experimentally ( fig . [ zerobias ] ( a ) ) . by separately contacting the two layers , the edge states can carry different currents , and the relative current @xmath66 is the exciton current . if we switch off the tlj ( by turning off the gate voltages ) , the two layers are decoupled and the voltage in each layer is proportional to its current . therefore @xmath67 . with the tlj the two layers are exchanged when the edge states meet the tlj , so that the two voltages are also reversed , and @xmath68 . the cases with and without tlj correspond to perfect andreev and perfect normal scattering of the exciton . to verify that the zero bias peak conductance observed from the perfect andreev " reflection experiment is really from a local parafermion zero mode , we can further consider the experimental setting in fig . [ zerobias ] ( b ) , which is an analog of using a scanning tunneling microscope ( stm ) to probe the majorana zero mode . by contacting source and drain to the two different layers and bringing the edge states close to the tlj , the edge states carry the exciton current @xmath69 , and play the role of an stm tip . it is well - known that the electron tunneling from an stm tip leads to a differential conductance proportional to the local electron density of states . by analogy , the tunneling conductance in our system measures the local density of states of the exciton , related to @xmath70 . @xmath71 annihilates an exciton in the upper edge of the tlj , or equivalently , tunnels a @xmath72 quasiparticle from layer @xmath73 to layer @xmath20 . we find @xmath74 has the following exponential behavior : @xmath75 , where @xmath76 is the correlation length of the gapped region ( with @xmath45 the gap opened at the tlj ) , and @xmath77 is a power - law function of @xmath78 . the exponential localization is remarkable , since part of the edge states are still gapless . experimental setups to measure zero bias interlayer quasiparticle conductance . ( a ) the `` andreev reflection '' process ( see text ) in which the exciton current @xmath79 is reversed across tlj . ( b ) use the edge states of the sample to measure the local quasiparticle density of states , which detect the parafermion zero modes at the ends of the tlj.(c ) edge theory picture of tlj in ( b ) . the coordinate @xmath80 increase from @xmath81 to @xmath82 following the direction indicated by orange , yellow , red and blue dots . the second layer ( blue ) is flipped in the right panel . dotted lines between counterpropagating edge states indicate backscattering , generating an energy gap . , width=336 ] to understand this , consider the edge theory description of the tlj in fig . [ zerobias ] ( b ) . in eq . ( [ h0 ] ) , ( [ ht ] ) , we used separate chiral boson fields @xmath83 to describe the two edges from each layer . here , for convenience we describe the whole edge of the first ( second ) layer by a single chiral boson @xmath84 . therefore @xmath85 each lives on a circle with perimeter @xmath86 ( with @xmath87 the length of the tlj ) . the edge of each layer is a chiral luttinger liquid , and it is convenient to group the two chiral luttinger liquids into a non - chiral luttinger liquid . for this purpose we define the spatial coordinate @xmath88 $ ] of the two edges with opposite chiralities ( fig . [ zerobias ] ( c ) ) . @xmath80 increases along the counter - clockwise direction for layer @xmath20 and clockwise direction for layer @xmath73 . with such a choice , @xmath89 and @xmath90 has opposite chirality ( defined with respect to @xmath80 ) , and the vertical back - scattering induced at the tlj occurs locally between @xmath91 and @xmath92 . the luttinger liquid lagrangian can be written as @xmath93.\end{aligned}\ ] ] with @xmath94 and where @xmath95 is a step function . the tunneling induces a gap in the region @xmath96 $ ] , and the region @xmath97 $ ] remains gapless . in this representation , the @xmath54 topological degenerate ground states are correctly given by the @xmath54 minima of the @xmath98 term , with @xmath99 pinned to the values @xmath100 . importantly , in this coordinate the exciton annihilation operator becomes nonlocal : @xmath101 . due to the non - local commutation relation of chiral boson fields , applying this operator to the ground state causes a domain wall in @xmath99 in the region @xmath102 . away from the defects at @xmath103 and @xmath104 , such a domain wall costs an energy of order the gap @xmath45 . therefore , the correlations of this operator must decay exponentially , away from the positions of the defects . at the positions of the defects , this operator keeps the system in the ground state manifold because it does not create a domain wall in @xmath99 in the gapped region . instead , it simply yields a unitary transformation among the ground states by globally flipping the value of @xmath99 in the gapped region . from this analysis we see that the zero bias peak in the exciton local density of state is indeed probing the topologically degenerate states . _ 2 . long - time current - noise cross correlation _ since the genons are non - abelian defects , it should be possible to detect them through interferometry . compared to the interferometry measurements proposed for detecting non - abelian anyons in non - abelian fqh states@xcite , an advantage of our current setup is that the non - abelian defects ( genons ) are imposed externally rather than being dynamical quasiparticle excitations , so that their position and number can be easily controlled . our setup ( fig . [ interferometer ] ) consists of two tljs and four quantum point contacts ( qpcs ) , created by either a top gate or a bottom gate . one pair of qpcs ( @xmath105 in fig . [ interferometer ] ) allows for weak quasiparticle tunneling across the normal region of the fluid , between the edges of the top layer . the other pair ( @xmath106 ) allows for weak quasiparticle tunneling across the tljs , from one edge of the top layer to the other edge of the bottom layer . it should be noticed that there are only two nontrivial interference loops formed by the four qpcs , the first one ( @xmath87 in fig . [ interferometer ] ) in the top layer including @xmath107 and @xmath108 , and the second one ( @xmath109 ) going through both layers including @xmath110 and @xmath111 . other interference paths such as the one enclosing @xmath107 and @xmath110 does not form a closed loop since the quasiparticle can not return to the same layer . the interference in these two loops @xmath112 measure the electric charge in the loops , which measure the topological qubit state of the genons @xmath113 and @xmath114 respectively . as a key consequence of the non - abelian nature of genons , the topological qubit degrees of freedom are non - local , and the topological charge carried by genons @xmath113 does not commute with that of @xmath114 . therefore non - local correlation between interference patterns in these two loops are induced due to the uncertainty principle . alternatively , the non - commutativity of the two quasiparticle paths @xmath87 and @xmath109 can also be understood from the geometrical picture shown in fig . [ doublestaircase ] , with the measured topological charge corresponding to @xmath49 , @xmath50 . qpc interferometer setup for detecting the non - abelian nature of genons . @xmath105 are the two qpc s in the first layer , and @xmath106 are the two inter - layer qpc s across the tljs . @xmath87 and @xmath109 indicated by thick dashed lines are the two interference loops . , width=336 ] to be more precise , the interference pattern can be observed by applying a voltage difference @xmath115 between the left and right - moving edges in both layers , and measuring the currents @xmath116 ( fig . [ interferometer ] ) . we find that the interlayer current noise cross - correlation @xmath117 has a dramatic feature : @xmath118 such a long time noise correlation , at finite temperatures , is a striking direct consequence of the non - abelian nature of the genons , which has never been proposed or observed before in an abelian state . intuitively , this is understood as follows . the current @xmath119 depends on the quasiparticle interference between the two qpcs @xmath105 . when a quasiparticle leaves the top layer through the qpc @xmath110 , its fractional statistics cause a phase shift in the interference of loop @xmath87 and therefore changes the value of @xmath119 . such a phase shift is permanent until the next quasi - particle tunneling occurs at @xmath110 . the converse is also true , that the quasiparticle tunneling at @xmath108 causes a permanent phase shift in @xmath120 . therefore , the fluctuation of the currents @xmath119 and @xmath121 have a long - time correlation for both @xmath122 and @xmath123 . it should be noticed that only having three qpc s @xmath124 is topologically equivalent to a setup proposed in @xcite , as a way to measure fractional statistics . ] can already induce a long - time correlation of @xmath125 , but only for @xmath122 . we present in the supplemental materials an explicit calculation of @xmath125 based on the edge theory ; the result , for @xmath126 and to lowest order in the tunneling strength at the qpc s , takes the form : @xmath127.\end{aligned}\ ] ] the dimensionless functions @xmath128 , whose explicit form is given in the supplemental materials , depend on the quasiparticle statistics ( they are non - zero only when @xmath129 ) and decay with the dimensionless ratios @xmath130 , and @xmath131 , which are optimally of order one , where @xmath78 is the distance scale between the qpcs . on much longer time scales , thermally excited mobile quasiparticles can decohere the state of the qubit , which will wash out the noise signature considered here . however this time scale is expected to be exponentially large in @xmath132 , ( _ ie @xmath133 ) , because the quasiparticle density is exponentially small in @xmath132 . _ * discussion * a longer term goal is to implement unitary gates on the topological qubit by braiding genons . this can be realized with further sets of gates to tune interactions between various zero modes localized at the genons.@xcite the computational power of genons in generic abelian states is stronger than that provided by majorana zero modes @xcite . the setup proposed here can be generalized to realize universal tqc in non - abelian states that by themselves would not be universal for tqc @xcite . examples include two layers of the non - abelian phase of kitaev s honeycomb model @xcite , or the odd - denominator bonderson - slingerland fqh states @xcite . _ acknowledgments we would like to thank david goldhaber - gordon , woo - won kang , charlie marcus and hari manoharan for helpful conversations . this work was supported by the simons foundation ( mb ) and david and lucile packard foundation ( xlq ) . _ 44 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , , , * * , ( ) . , _ _ ( , , ) . , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , , , , , , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , * * , ( ) . , , ( ) , . , * * , ( ) . , , ( ) , . , , , , * * , ( ) . , * * , ( ) . ( ) , . , , ( ) , . , , ( ) , . , , , , , * * , ( pages ) ( ) . , , , , , , , , * * , ( ) . , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , * * , ( ) . , , , , * * , ( ) . , , , , , * * , ( ) . , , , , , * * , ( ) . , * * , ( ) . , ( ) . , * * , ( ) . , * * , ( ) . , ( ) . , , , * * , ( ) . , , , , , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , ( ) . , , , * * , ( ) . , , , , , * * , ( ) . , , , , * * , ( ) . , * * , ( ) .

the multiplicity function ( hereafter mf ) is a powerful tool to test the various cosmological scenarios ( combes & boiss 1991 ) and provides an independent way to measure the index @xmath1 of the initial perturbation spectrum @xmath1 at the era of baryonic recombination ( turner , gott iii 1976 ) . + the cronario project is a joint enterprise among caltech and the astronomical observatories of napoli , roma and rio de janeiro , aimed to produce the first general catalogue of all objects visible on the dposs ( digitised palomar sky survey ) . the final palomar - norris catalogue will include astrometric , photometric ( in the three gunn - thuan bands g , r and i ) and rough morphological information for an estimated @xmath2 stars and @xmath3 galaxies . more than @xmath4 of the catalogues are already available and are currently being used for many scientific applications . in order to compile a catalog of candidate groups of galaxies in absence of redshift information , we implemented a slightly modified version of the van albada algorithm ( soares 1989 ) . + this algorithm makes use of apparent magnitude and projected position in the sky only , and gives , for each pair of adjacent galaxies , their probability of being physically related . + assuming a poisson statistic , the probability that the angular distance of a fixed galaxy to its nearest non - physical companion lies between @xmath5 and @xmath6 is given by : @xmath7 \cdot 2 \pi \theta n d\theta.\ ] ] the introduction of an adimensional distance @xmath8 ( defined as the ratio of the observed to the expected mean distance to the nearest neighbour ) allows to combine the angular separation of different pairs into a single distribution ( fig . [ fig : poisson ] ) removing the effects of clustering in the background . + = 9.0 cm in order to test the accuracy ( lost groups ) and the reliability ( spurious groups ) of the algorithm , we tested it on @xmath9 realistically simulated sky fields . + we first produced the galaxy background by assuming uniform distribution and the field luminosity function given by metcalfe et al . 1994 ; then we added simulated groups of galaxies according to the multiplicity function by turner & gott iii 1976 , with redshift computed according to : @xmath10\ ] ] and absolute magnitude of the brightest galaxy in the group taken from the cumulative luminosity function for groups of galaxies : @xmath11^{\alpha+1 } \cdot exp \left [ -10^{\left ( 0.4 \left ( m^*-m \right ) \right ) } \right ] dm\ ] ] where @xmath12 and @xmath13 . other parameters which were varied in the course of the simulations were : * the diameter of the group inside a gaussian distribution centered at : @xmath14 ; * the maximum possible redshift for a group , which were assumed to fall in the range : @xmath15 . different simulations were performed in order to optimize the lower limit of the probability for which two galaxies were considered as physical companions by the algorithm . the optimal value , i d est the value ensuring the best compromise between accuracy and reliability ( fig . [ fig : prob ] ) , turned out to be @xmath16 . + with this choice , the modified van albada algorithm succeds in retrieving , from the simulated catalogues , more than @xmath17 of the groups having more than three components . + = 12.0 cm the figure ( fig . [ fig : disc ] ) gives the mf obtained from @xmath18 cronario plates ( subtending a total solid angle of @xmath19 ) . in the same figure we also compare it with six other mfs taken from literature : 1 . * mf derived from magnitude limited surveys making use of redshift information : * * garcia * ( garcia et al . 1993 ) , * cfa * ( geller & huchra 1983 ) , * northcfa * ( ramella , pisani 1997 ) , * esp * ( ramella et al . 1998 ) 2 . * mf derived from diameter limited survey : maia * ( maia & da costa 1989 ) 3 . * mf derived from limited magnitude survey without redshift information : turner * ( turner & gott 1976 ) agreement is found between our fm , obtained with our algorithm without any tridimensional data , and the ones in literature using redshift information . therefore , with our algorithm , even in absence of redshift information , it is possible to explore wide areas of the sky and to retrieve groups ( @xmath20 ) with an efficiency comparable to that of 3-d surveys . + = 6.5 cm for the groups with at least one redshift available , mf ( as a function of luminosity ) was calculated . with a fit of the observed data to the theoretical function @xmath21 we obtained : @xmath22 + this allows an estimate of the spectral index @xmath1 : @xmath23

the projected multiplicity function of galaxies gives ( per square degree ) the density of galaxy aggregates formed by n members . we use the cronario matched catalogues to derive the low n ( @xmath0 ) multiplicity function from d - poss data . the van albada ( 1982 ) algorithm was implemented and used to identify candidate groups in the cronario catalogues . in absence of redshift surveys complete down to the magnitude limit of the dposs material and covering a wide enough area , the performances of the algorithm were tested on realistically simulated catalogues . + the application to a set of 13 cronario catalogues allowed us to derive a multiplicity function as accurate as those available in literature obtained from redshift surveys . # 1 # 2 _ mem . soc . astron . it . _ * # 1 * , # 2 # 1 # 2 _ the messenger _ * # 1 * , # 2 # 1 # 2 _ astron . nach . _ * # 1 * , # 2 # 1 # 2 _ astron . astrophys . _ * # 1 * , # 2 # 1 # 2 _ astron . astrophys . lett . _ * # 1 * , l#2 # 1 # 2 _ astron . astrophys . rev . _ * # 1 * , # 2 # 1 # 2 _ astron . astrophys . suppl . ser . _ * # 1 * , # 2 # 1 # 2 _ astron . j. _ * # 1 * , # 2 # 1 # 2 _ ann . rev . astron . astrophys . _ * # 1 * , # 2 # 1 # 2 _ astrophys . j. _ * # 1 * , # 2 # 1 # 2 _ astrophys . j. lett . _ * # 1 * , l#2 # 1 # 2 _ astrophys . j. suppl . _ * # 1 * , # 2 # 1 # 2 _ astrophys . space sci . _ * # 1 * , # 2 # 1 # 2 _ adv . space res . _ * # 1 * , # 2 # 1 # 2 _ bull . astron . inst . czechosl . _ * # 1 * , # 2 # 1 # 2 _ j. quant . spectrosc . . transfer _ * # 1 * , # 2 # 1 # 2 _ mon . not . r. astr . soc . _ * # 1 * , # 2 # 1 # 2 _ mem . r. astr . soc . _ * # 1 * , # 2 # 1 # 2 _ phys . lett . rev . _ * # 1 * , # 2 # 1 # 2 _ publ . astron . soc . japan _ * # 1 * , # 2 # 1 # 2 _ publ . astr . soc . pacific _ * # 1 * , # 2 # 1 # 2 _ nature _ * # 1 * , # 2 epsf.sty

as is well known , most materials , regardless how pure they may seem at the macroscopic level , have a rather complicated microscopic structure . it may then come as a surprise that physical phenomena such as heat or electric conduction are described so well using differential equations with smooth , sometimes even constant , coefficients . an explanation has been offered by homogenization theory : rapid oscillations at the microscopic level average out , or homogenize , at the macroscopic scale . however , this does not mean that the microscopic structure is simply washed out . indeed , while it disappears from the structure of the resulting equations , it remains embedded in the values of effective material constants , e.g. , the coefficients . an illustrative example of a homogenization problem is that of effective conductance . we will formulate an instance of this problem in the setting of resistor networks . let @xmath0 denote the @xmath5-dimensional hypercubic lattice and suppose that each unordered nearest - neighbor edge @xmath6 is assigned a value @xmath7 called the _ conductance _ of @xmath6 . for any @xmath8 , let @xmath9 be the edges with at least one endpoint in @xmath10 . given an @xmath11 and a finite @xmath8 , let @xmath12 ^ 2,\ ] ] where each pair @xmath13 is counted only once . this is the electrostatic ( dirichlet ) energy for the potential @xmath14 with dirichlet boundary condition on the boundary vertices of @xmath10 . consider now the square box @xmath15 . a quantity of prime interest for us is the _ effective conductance _ , @xmath16 where @xmath17 and where @xmath18 are those vertices outside @xmath10 that have an edge into @xmath10 . by kirchhoff s and ohm s laws ( see , e.g. , doyle and snell @xcite ) , @xmath19 is the total electric current flowing through the network when the boundary vertices are kept at voltage @xmath20 . for homogeneous resistor networks , i.e. , when @xmath21 for all @xmath6 , the infimum is achieved by @xmath22 and so @xmath23 . a question of ( reasonably ) practical interest is then what happens when the conductances @xmath24 are no longer constant , but remain close to a constant . in particular , we may assume that they are uniformly elliptic , i.e. , @xmath25 a comparison of @xmath26 with these @xmath24 s and the homogeneous case shows that @xmath27 is still of the order of @xmath28 . moreover , thanks to the choice of the linear boundary condition , by subadditivity arguments the limit @xmath29 exists almost surely for any ergodic distribution of the conductances . the problem left to resolve is thus a computation of the limit value . although @xmath30 can be computed explicitly only in a handful of ( mostly periodic ) cases , it can be characterized in large generality : suppose that @xmath31 is a sample from a shift - ergodic law @xmath32 on the product space @xmath33 $ ] indexed by edges of @xmath0 . ( technically , @xmath34 is the coordinate projection on edge @xmath6 . ) as is well known ( papanicolaou and varadhan @xcite , kozlov @xcite and knnermann @xcite and the book jikov , kozlov and oleinik @xcite ) , @xmath35 here @xmath36 is expectation with respect to @xmath32 , @xmath37 are the unit coordinate vectors in @xmath38 and @xmath39 is the gradient of @xmath40 in direction of @xmath41 with @xmath42 denoting the shift by @xmath43 ; i.e. , the map on @xmath44 such that @xmath45 . the expression in can be interpreted as the dirichlet energy density with the spatial average naturally replaced by the ensemble average . once the ( deterministic ) leading - order of @xmath27 has been identified , the next natural question is that of fluctuations . it is obvious e.g. , by checking the explicitly computable @xmath46 case that no universal limit law can be expected for general conductance distributions , but progress could perhaps be made for the ( physically most appealing ) case of i.i.d . conductances . however , even here establishing just the order of magnitude of the fluctuations turned out to be an arduous task . indeed , more than a decade ago wehr @xcite showed that @xmath47 for some @xmath48 but a corresponding upper bound has been furnished only recently by gloria and otto @xcite . both of these results contain important technical caveats : wehr requires continuously distributed @xmath24 s while gloria and otto express their results under a `` massive '' cutoff ( at least in dimension 2 ) . gloria and otto drew important ideas from an earlier unpublished note by naddaf and spencer @xcite where ( optimal ) upper bounds on the variance are derived for certain correlated conductance laws . the main tool of @xcite is the _ meyers estimate _ ( cf meyers @xcite ) , to be used heavily in the present note as well . other noteworthy earlier derivations of ( suboptimal ) variance upper bounds include an old paper by yurinskii @xcite and a more recent paper by benjamini and rossignol @xcite . closely related to these estimates are recent derivations of quantitative central limit theorems for random walk among random conductances and approximations of the limiting diffusivity matrix , e.g. , caputo and ioffe @xcite , bourgeat and piatnitski @xcite , boivin @xcite , mourrat @xcite , gloria and mourrat @xcite , etc . incidentally , the meyers estimate is also the key tool in @xcite . the goal of the present note is to prove that , for i.i.d . conductances which are ( deterministically ) not too far from a constant , the asymptotic law of @xmath27 is in fact gaussian . let @xmath49 denote the normal random variable with mean @xmath50 and variance @xmath51 . then we have : [ thm1 ] suppose the conductances @xmath24 are i.i.d . for each @xmath52 , there is @xmath53 such that the following holds : if is satisfied @xmath32-a.s . with this @xmath54 , then for each @xmath17 there is @xmath55 such that @xmath56 whenever the conductance law is non - degenerate we have @xmath57 for all @xmath58 . the proof also immediately yields : [ cor1.2 ] under the conditions of theorem [ thm1 ] , @xmath59 where @xmath60 is as in . a few remarks are in order : \(1 ) notice that does not give us much information on the `` order expansion '' of @xmath61 . indeed , we know that @xmath62 is to the leading order equal to @xmath63 but when this order is subtracted , the next - order term is ( presumably ) of boundary size . in @xmath64 , this is still larger than the typical size of the fluctuations . ( this is referred to as the systematic error in @xcite . ) notwithstanding , what does tell us is the character of the leading order _ random _ term . \(2 ) there is in fact a formula for @xmath60 , see theorem [ thm2 ] below , which also shows that @xmath65 is a bi - quadratic ( and thus smooth ) function of @xmath66 . however , the formula involves complicated conditioning and does not seem very useful for practical computations . \(3 ) there is no restriction on the single - conductance law other than . in particular , this law can have a non - absolutely continuous part including atoms . certain technical problems do arise at this level of generality ; see section [ sec2.5 ] which , we believe , is of independent interest . we prove theorem [ thm1 ] by reducing it to the martingale central limit theorem . there are two main technical ingredients : homogenization theory ( which enables a stationary martingale approximation of @xmath27 ) and analytical estimates for finite - volume harmonic coordinates ( by which we control the errors in the martingale approximation ) . the restrictions to rectangular boxes , linear boundary conditions and small ellipticity contrasts permit us to encapsulate the analytical input into a single step , the meyers estimate , cf proposition [ prop : meyers estimate ] and theorem [ lpbound - k ] . these restrictions can be relaxed but not without additional arguments not all of which have been handled satisfactorily at this time . these are deferred to a follow - up paper . we remark that two recent preprints have been brought to our attention at the time this work was first announced in conference talks . first , nolen @xcite has established a normal approximation to the effective conductance defined over a periodic environment , in the limit when the period tends to infinity . second , in a preprint that was posted at the time of writing the present note , rossignol @xcite formulates and proves a central limit law for the _ effective resistance _ for the corresponding problem on a torus . nolen defines the problem over continuum , albeit with a rather strong assumption on an underlying gaussian i.i.d . structure . rossignol s setting is based on minimizing the electrostatic energy over currents ( rather than potentials ) subject to a restriction on the total current flowing around the torus . by a well known reciprocity relation between effective conductance and resistance , these papers appear to address similar problems . the present paper differs from both nolen @xcite and rossignol @xcite primarily in its emphasis on fixed ( dirichlet ) , as opposed to periodic , boundary conditions . indeed , a majority of our technical work is aimed at controlling the resulting boundary effects . also the way a gaussian limit law is established is quite different : nolen appeals to stein s method , rossignol uses noise - sensitivity tools while we invoke the martingale central limit theorem . a notable deficiency of the present paper compared to @xcite and @xcite is the limitation on ellipticity contrast . nolen overcomes this by an appeal to gloria and otto @xcite , although this ultimately precludes the most interesting conclusion ( namely , the clt without `` massive '' cutoff ) in @xmath67 . approach appears to work seamlessly for all elliptic product laws thanks to `` enhanced '' form of averaging coming from the periodic boundary condition . while the gloria - otto method can perhaps be adapted to our situation as well , just as for nolen @xcite it fails to deliver the desired conclusion in @xmath67 . the issue is that the method yields bounds on the moments of the corrector , which diverge in @xmath67 , while we need only moments of the gradients of the corrector . ( _ update in revised version _ : this issue has now been overcome in gloria , neukamm and otto @xcite , albeit only in either infinite volume or for periodic boundary conditions . ) notwithstanding , our point of view is that the moment bounds seem to be a separate technical matter , and so , for the present paper , we decided to sacrifice on generality of the distribution and derived the clt only in the simplest , albeit still physically interesting , case . here we discuss the strategy of the proof of theorem [ thm1 ] and state its principal ingredients in the form of suitable propositions . the actual proofs begin in section [ sec3 ] . -1 mm a standard way to control fluctuations of a function of i.i.d . random variables is by way of a _ martingale approximation_. let us order the random variables @xmath68 in any ( for now ) convenient way and let @xmath69 to be the @xmath70-algebra generated by the first @xmath71 of them . ( since we only aim at a distributional convergence , the @xmath70-algebras may depend on @xmath72 . ) then @xmath73 where @xmath74 obviously , the quantity @xmath75 is a martingale increment . in order to show distributional convergence to @xmath76 , it suffices to verify the ( lindenberg - feller - type ) conditions of the martingale central limit theorem due to brown @xcite : 1 . there exists @xmath77 such that @xmath78 in probability , and 2 . for each @xmath79 , @xmath80 in probability . the sums on the left suggest invoking the spatial ergodic theorem , but for that we would need to ensure that the individual terms in the sum are ( at least approximated by ) functions that are stationary with respect to shifts of @xmath0 . this necessitates the following additional input : 1 . a specific choice of the ordering of the edges , and 2 . a more explicit representation for @xmath75 . we will now discuss various aspects of these in more detail . recall that @xmath81 denotes the set of all ( unordered ) edges in @xmath0 . we will order @xmath81 as follows : let @xmath82 denote the lexicographic ordering of the vertices of @xmath0 . explicitly , for @xmath83 and @xmath84 we have @xmath85 if either @xmath86 or @xmath87 and there exists @xmath88 such that @xmath89 for all @xmath90 and @xmath91 . we will write @xmath92 if @xmath87 and @xmath85 . for the purpose of defining a stationary ordering of the edges , and also easier notation in some calculations that are to follow , we now identify @xmath81 with the set of pairs @xmath93 , where @xmath41 and @xmath88 , so that @xmath93 corresponds to the edge between the vertices @xmath43 and @xmath94 . we will then write @xmath95 again , @xmath96 if @xmath97 but @xmath98 . it is easy to check that @xmath82 is a complete order on @xmath81 . a key fact about this ordering is its stationarity with respect to shifts : [ lemma-2.3 ] if @xmath97 then also @xmath99 for all @xmath100 . proof this is a trivial consequence of the definition . now we proceed to identify the sigma algebras @xmath101 in the martingale representation above . recall that @xmath33 $ ] denotes the set of conductance configurations satisfying . writing @xmath102 for elements of @xmath44 we use @xmath31 , for @xmath103 , to denote the coordinate projection corresponding to edge @xmath6 . given @xmath104 , set @xmath105 and let @xmath106 be the enumeration of @xmath107 induced by the ordering of edges @xmath82 defined above . then we set @xmath108 with @xmath109 by definition @xmath110 is independent of the edges in @xmath107 while @xmath111 determines the entire configuration in @xmath107 . note also that @xmath112 includes information about edges that are not in @xmath107 . this will be of importance once we replace @xmath75 by a random variable that depends on all of @xmath102 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ figure 1 . the set of edges ( drawn in bold ) in @xmath107 that determine the events in @xmath112 . here @xmath113 , @xmath114 and @xmath115 . in our ordering of edges on @xmath116 , this corresponds to the edges up to and including @xmath93 for @xmath117 and @xmath118 . [ fig1 ] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ having addressed the ordering of the edges , and thus the definition of the @xmath70-algebras @xmath112 , we now proceed to derive a more explicit form of the quantity @xmath75 from . given @xmath119 , define the operator @xmath120 on ( @xmath121 or @xmath38-valued ) functions on the lattice via @xmath122.\ ] ] this is an elliptic finite - difference operator a random laplacian that arises as the generator of the random walk among random conductances @xmath123 ( see , e.g. , biskup @xcite for a review of these connections ) . the existence / uniqueness for the associated dirichlet problem implies that for any finite @xmath8 there is a unique @xmath124 such that @xmath125 obeys @xmath126 & \psi_\lambda(\omega , x)=x , & \qquad & x\in\partial\lambda . \end{alignedat}\right.\ ] ] it is then easily checked that @xmath127 is the unique minimizer of @xmath128 over all functions @xmath14 with the boundary values @xmath129 for @xmath130 . in particular , we have @xmath131 for all @xmath17 . the function @xmath125 will sometimes be referred to as a _ finite - volume harmonic coordinate_. ( the first line in justifies this term . ) the minimum value @xmath132 is a non - decreasing , continuous and concave function of @xmath133 . thanks to the uniqueness of the solution to , @xmath132 is also continuously differentiable in @xmath24 s and it is easy to see , keeping in mind again that @xmath134 is the minimizer and using the euler - lagrange equation for the minimization problem , that @xmath135 ^ 2,\qquad \langle x , y\rangle\in{\mathbb b}(\lambda).\ ] ] this relation is of fundamental importance for what is to come . abusing the notation slightly , let @xmath136 , with @xmath137 , denote the components of the configuration @xmath102 over @xmath9 labeled in the order induced by @xmath82 defined above . let @xmath138 mark explicitly the dependence of the right - hand side on these variables . the product structure of the underlying probability measure then allows us to give a more explicit expression for the increment @xmath139 : @xmath140 \\ & = \int { \mathbb p}({\text{\rm d}\mkern0.5mu}\omega_k')\dots{\mathbb p}({\text{\rm d}\mkern0.5mu}\omega_n')\int_{\omega_k'}^{\omega_k}{\text{\rm d}\mkern0.5mu}\tilde\omega_k\ , \frac\partial{\partial\tilde\omega_k}q(\omega_1,\dots,\omega_{k-1},\tilde\omega_k,\omega_{k+1}',\dots,\omega_n ' ) , \end{aligned}\ ] ] with the inner integral in riemann sense . a key point is that the last partial derivative is ( modulo notational changes ) given by , i.e. , @xmath75 is the modulus - squared of the gradient of @xmath134 over the @xmath71-th edge in @xmath9 integrated over part of the variables . to see that @xmath75 is a martingale increment note that the riemann integral changes sign when its limits are interchanged . in order to apply the spatial ergodic theorem to the sums on the left of ( [ e:2.4a][e:2.5a ] ) , we will substitute for @xmath75 a quantity that is stationary with respect to the shifts of @xmath0 . this will be achieved by replacing the discrete gradient of @xmath141 which by enters as the partial derivative of @xmath142 in the formula for @xmath75 by the gradient of its infinite - volume counterpart , to be denoted by @xmath143 . the existence and properties of the latter object are standard : [ prop - corrector ] suppose the law of the conductances is ergodic with respect to the shifts of @xmath0 and assume for some @xmath144 . then there is a function @xmath145 such that 1 . ( @xmath143 is @xmath120-harmonic ) @xmath146 for all @xmath43 and @xmath32-a.e . ( @xmath143 is shift covariant ) for @xmath32-a.e . @xmath102 we have @xmath147 and @xmath148 3 . ( @xmath143 is square integrable ) @xmath149 4 . ( @xmath143 is approximately linear ) the corrector @xmath150 satisfies @xmath151 proof properties ( 1 - 3 ) are standard and follow directly from the construction of @xmath143 ( which is done , essentially , by showing that a minimizing sequence in converges in a suitable @xmath152-sense ; see , e.g. , biskup ( * ? ? ? * section 3.2 ) for a recent account of this ) . as to ( 4 ) , a moment s thought reveals that it suffices to show this for @xmath43 of the form @xmath153 , where @xmath154 . this follows from the mean ergodic theorem , similarly as in ( * ? ? ? * lemma 4.8 ) . the replacement of ( the gradients of ) @xmath141 by @xmath143 necessitates developing means to quantify the resulting error . for this we introduce an @xmath155-norm on functions @xmath156 by the usual formula @xmath157 analogously , we also introduce a norm on functions @xmath158 by @xmath159 here we introduced the symbol @xmath160 for an @xmath38-valued function whose @xmath161-th component at @xmath43 is given by @xmath162 abusing our earlier use of this notation . it is reasonably well known , albeit perhaps not written down explicitly anywhere , that the gradients of @xmath141 and @xmath143 are close in @xmath163-norm ( see , however , proposition 3.1 of caputo and ioffe @xcite for a torus version of this statement ) . [ prop - approx - corrector - l2 ] suppose the law @xmath32 on conductances @xmath164 is ergodic with respect to shifts of @xmath0 and obeys for some @xmath144 . then @xmath165 as we will elaborate on later ( see remark [ r : representation ] ) , this is exactly what is needed to establish the representation for the limit value @xmath30 of the sequence @xmath166 . however , in order to validate the conditions ( [ e:2.4a][e:2.5a ] ) of the martingale central limit theorem , more than just square integrability is required . for this we state and prove : [ prop : meyers estimate ] suppose @xmath32 is ergodic with respect to shifts . for each @xmath52 , there is @xmath53 such that if holds @xmath32-a.s . with this @xmath54 , then for some @xmath167 , @xmath168 and @xmath169 proposition [ prop : meyers estimate ] is the sole reason for our restriction on ellipticity contrast . we believe that , on the basis of the technology put forward in gloria and otto @xcite , no such restriction should be needed . to attest this we note that versions of the above bounds actually hold pointwise for a.e . @xmath119 satisfying ; i.e. , for norms without the expectation @xmath36 . in addition , from ( * ? ? ? * proposition 2.1 ) we in fact know for all @xmath170 when @xmath64 . _ update in revised version _ : we note that is now known in all @xmath171 from gloria , neukamm and otto @xcite . unfortunately , this does not apply to . a torus version of proposition [ prop : meyers estimate ] appeared in theorem 4.1 of caputo and ioffe @xcite . unfortunately , a direct attempt at the substitution of ( the gradients of ) @xmath141 by @xmath143 in reveals another technical obstacle : as relies on the fundamental theorem of calculus , the replacement of @xmath141 by @xmath143 requires the latter function to be defined for @xmath102 that may lie outside of the support of @xmath32 . this is a problem because @xmath143 is generally determined by conditions ( 1 - 4 ) in proposition [ prop - corrector ] only on a set of full @xmath32-measure . imposing additional assumptions on @xmath32 namely , that the single - conductance distribution is supported on an interval with a bounded and non - vanishing density would allow us to replace the lebesgue integral in by an integral with respect to @xmath172 and thus eliminate this problem . notwithstanding , we can do much better by invoking a rank - one perturbation argument which we describe next . fix an index @xmath88 and recall the notation @xmath162 . for a vertex @xmath41 and a finite set @xmath8 satisfying @xmath173 or @xmath174 , let @xmath175 be defined by @xmath176 obviously , @xmath177 is continuous in @xmath178 for @xmath179 and it is bounded away from @xmath180 and infinity . ( in section [ sec - perturbed - coordinate ] we will see that @xmath177 is in fact a double gradient of the green function for operator @xmath120 on @xmath10 . ) note that and ask us to understand how @xmath181 changes when the coordinate of @xmath102 over @xmath182 is perturbed . this change takes a multiplicative form : [ prop - perturbed - corrector ] let @xmath8 be finite and @xmath183 be nearest neighbors ; @xmath184 for some @xmath88 . for any @xmath185 that agree everywhere except at edge @xmath186 , @xmath187\nabla_i\psi_\lambda(\omega , x).\ ] ] for the prefactor we get the alternative expressions @xmath188 in particular , the factor @xmath189 is bounded away from 0 and @xmath190 uniformly in @xmath119 and @xmath8 . it is worthy a note that is a direct consequence of a rank - one perturbation formula for the green function ; cf lemma [ lemma-5.1 ] . incidentally , such formulas have proved extremely useful in the analysis of random schrdinger operators , including those associated with operator @xmath191 ( aizenman and molchanov @xcite ) . what matters for us is the @xmath192-limit of the expression in can be controlled uniformly in @xmath119 : [ prop-2.8 ] recall that @xmath193^{{\mathbb b}({\mathbb z}^d)}$ ] for some @xmath144 . then @xmath194 is non - decreasing and bounded away from zero and infinity uniformly in @xmath8 and @xmath119 . moreover , for all @xmath119 and all @xmath41 the limit @xmath195 exists and satisfies @xmath196 where @xmath197 . in particular , @xmath198 is monotone and continuous in the product topology on @xmath44 . finally , @xmath199 for @xmath179 is continuous and bounded and @xmath200 is stationary in the sense that @xmath201 holds for all @xmath119 and all @xmath202 . before we wrap up the outline of the proof of theorem [ thm1 ] , let us formulate a representation for the limiting variance @xmath60 from theorem [ thm1 ] : for @xmath41 and @xmath88 , let @xmath203 denote the edge corresponding to the pair @xmath93 and let @xmath204 ^ 2,\ ] ] where @xmath205 is the configuration equal to @xmath102 except at @xmath203 , where it equals @xmath206 . define the matrix @xmath207 by the quadratic form @xmath208 where @xmath93 represents the edge @xmath182 and @xmath17 . then we have : [ thm2 ] under the assumptions of theorem [ thm1 ] , the matrix elements of @xmath209 are square integrable . in particular , @xmath60 from theorem [ thm1 ] is given by @xmath210 as an inspection of reveals , the limiting variance is thus a bi - quadratic form in @xmath66 . although concisely written , the expression is not very useful from the practical point of view ; particularly , due to the unwieldy conditioning in . notwithstanding , a slightly more explicit form can be obtained by performing the inner integral in with the help of : @xmath211 where , as before , @xmath212 agrees with @xmath102 except at @xmath179 . note that this implies , at least formally , that @xmath213 and so @xmath214 i.e. , that @xmath215 is a martingale increment . a question of interest is whether an expression can be found for @xmath60 that is more amenable to computations . since @xmath216 is quadratic in @xmath66 , and thus linear in @xmath217 , theorem [ thm2 ] and ( a version of ) the cramr - wold device imply that , as @xmath218 , the joint law of @xmath219 tends to a multivariate gaussian @xmath220 with @xmath221 where @xmath222 is as in . naturally , @xmath223 is a quadratic form as well . the proofs ( and the rest of the paper ) are organized as follows . in section [ sec3 ] we assemble the ingredients following the steps outlined in the present section into the proofs of theorems [ thm1 ] and [ thm2 ] and corollary [ cor1.2 ] . in section [ sec - meyers ] we then show that the finite - volume harmonic coordinate approximates its full lattice counterpart in an @xmath152-sense as stated in proposition [ prop - approx - corrector - l2 ] and establish the meyers estimate from proposition [ prop : meyers estimate ] . a key technical tool is the caldern - zygmund regularity theory and a uniform bound on the triple gradient of the green function of the simple random walk in finite boxes . finally , in section [ sec - perturbed - coordinate ] , we prove propositions [ prop - perturbed - corrector ] and [ prop-2.8 ] dealing with the harmonic coordinate over environments perturbed at a single edge . in this section we verify the conditions ( [ e:2.4a][e:2.5a ] ) of the martingale central limit theorem and thus prove theorems [ thm1 ] and [ thm2 ] . all derivations are conditional on propositions [ prop - approx - corrector - l2][prop-2.8 ] the proofs of which are postponed to later sections . throughout we will make use of the following simple but useful consequence of hlder s inequality : [ lemma : p_p ] for any @xmath224 , @xmath225 and @xmath226 , @xmath227 proof apply hlder s inequality to the function @xmath228 . assume now the setting developed in section [ sec2 ] ; in particular , the ordering of edges and sigma - algebras @xmath112 from section [ sec2.2 ] and the martingale increment @xmath75 from and its representation from section [ sec2.3 ] . in analogy with equation , we also define @xmath229 ^ 2,\ ] ] where @xmath179 and @xmath205 is the configuration equal to @xmath102 except at @xmath203 , where it equals @xmath206 . by proposition [ prop - perturbed - corrector ] , we may write the martingale increment @xmath75 as @xmath230 where @xmath231 and @xmath232 are the vertex and the edge direction corresponding to @xmath233 , i.e. , @xmath234 . ( a representation similar to is possible here as well . ) recall the notation for @xmath209 from and note that this is well defined and finite @xmath32-a.s . thanks to the estimates ( [ eqn - approx - corrector - lp-1][eqn - approx - corrector - lp-2 ] ) as well as boundedness of @xmath235 . note the dependence of @xmath75 on @xmath72 . [ prop - mclt1 ] assume that the premises ( and thus conclusions ) of propositions [ prop - approx - corrector - l2][prop-2.8 ] hold . then @xmath236 for all @xmath71 and @xmath237 in @xmath32-probability and @xmath238 . proof fix @xmath17 . thanks to lemma [ lemma-2.3 ] and proposition [ prop - corrector](2 ) , for each @xmath88 , the collection of conditional expectations @xmath239 is stationary with respect to the shifts on @xmath0 and , by proposition [ prop : meyers estimate ] , bounded in @xmath238 . labeling the edges in @xmath107 according to the order @xmath82 , the spatial ergodic theorem yields @xmath240 with the limit @xmath32-a.s . and in @xmath238 . to see how this relates to our claim , abbreviate @xmath241 and denote @xmath242 ^ 2- { \mathbb e}\big[b_k\,\big|{\mathcal f}_{k}\big]^2 \big|{\mathcal f}_{k-1}\big].\end{aligned}\ ] ] by we have @xmath243 , while reads @xmath244 . hence , as soon as we show that @xmath245 the claim will follow . the proof of will proceed by estimating @xmath246 which will involve applications of the cauchy - schwarz inequality ( in order to separate terms ) and jensen s inequality ( in order to eliminate conditional expectations ) . first we note @xmath247\bigr)^{{\mathchoice{\myffrac{1}{2 } in \scriptstyle } { \myffrac{1}{2 } in \scriptstyle } { \myffrac{1}{2 } in \scriptscriptstyle } { \myffrac{1}{2 } in \scriptscriptstyle } } } \bigl({\mathbb e}\bigl[(a_k+b_k)^2\bigr]\bigr)^{{\mathchoice{\myffrac{1}{2 } in \scriptstyle } { \myffrac{1}{2 } in \scriptstyle } { \myffrac{1}{2 } in \scriptscriptstyle } { \myffrac{1}{2 } in \scriptscriptstyle } } } .\ ] ] writing @xmath248 and noting @xmath249 tells us @xmath250 \le2{\mathbb e}\bigl[(a_k - b_k)^2\bigr]+8{\mathbb e}\bigl(b_k^2\bigr).\ ] ] summing over @xmath71 and applying the cauchy - schwarz inequality , we find that @xmath251 where @xmath252\quad\text{and}\quad \beta:=\frac1{|\lambda_l|}\sum_{k=1}^{|{\mathbb b}(\lambda_l)|}{\mathbb e}(b_k^2).\ ] ] by inspection of we now observe that it suffices to show that @xmath253 stays bounded while @xmath254 tends to zero in the limit @xmath218 . the boundedness of @xmath253 follows from and the fact that @xmath255 is bounded ; indeed , these yield @xmath256 uniformly in @xmath71 and @xmath72 . concerning the terms constituting @xmath254 , using @xmath257 we first separate terms as @xmath258\le2{\mathbb e}\bigl(\bigl|h_{\lambda_l}(\cdot , x_k , i_k)\bigr|^2\bigl|\,|\nabla_{i_k}(t\cdot\psi_{\lambda_l})(\cdot , x_k)|^2-|\nabla_{i_k}(t\cdot\psi)(\cdot , x_k))|^2\bigr|^2\bigr ) \\ + \ , 2{\mathbb e}\bigl(\bigl|h_{\lambda_l}(\cdot , x_k , i_k)-h(\cdot , x_k , i_k)\bigr|^2\bigl|\nabla_{i_k}(t\cdot\psi)(\cdot , x_k)\bigr|^4\bigr ) . \qquad\end{gathered}\ ] ] since @xmath259 is uniformly bounded , the average over @xmath71 of the first term is bounded by a constant times the product of @xmath260 and @xmath261 . the latter tends to zero as @xmath218 by proposition [ prop : meyers estimate ] , proposition [ prop - approx - corrector - l2 ] and lemma [ lemma : p_p ] ( with the choices @xmath262 and @xmath263 but sufficiently close to @xmath264 ) . for the second term in we pick @xmath167 and use hlder s inequality to get @xmath265 where @xmath142 satisfies @xmath266 . the norm of @xmath267 is again bounded by proposition [ prop : meyers estimate ] as long as @xmath268 is sufficiently close to @xmath264 ; to apply , we need to invoke the stationarity of @xmath269 to realize @xmath270 . for the second term in we first need to show that for each @xmath79 there is @xmath271 so that for all @xmath119 , @xmath272 for this we use that , thanks to , , and the monotonicity of @xmath273 , @xmath274 for some constant @xmath275 . to estimate the right - hand side , we invoke stationarity with respect to shifts and note that whenever @xmath276 , we have @xmath277 then follows from and the fact that the difference on the right - hand side of converges to zero uniformly in @xmath119 . ( specifically , we apply dini s theorem for uniformity : @xmath44 is compact in the product topology by tychonoff s theorem , @xmath278 is a non - decreasing sequence of continuous functions and the limit @xmath279 is continuous as well . ) we now bound the last term in as follows . the terms for which @xmath231 is at least @xmath280 steps away from @xmath281 are bounded by @xmath282 thanks to ; the sum over the remaining terms is of order @xmath283 thanks to the uniform boundedness of @xmath284 . hence , in the limit @xmath218 , the second term in is of order @xmath285 ; taking @xmath286 shows that @xmath254 tends to zero as @xmath218 . invoking , this finishes the proof of and the whole claim . [ prop - mclt2 ] assume that the premises ( and thus conclusions ) of propositions [ prop - approx - corrector - l2][prop-2.8 ] hold . then for each @xmath79 , @xmath287 in @xmath32-probability . proof this could be proved by strengthening a bit the statement of proposition [ prop - mclt1 ] ( from squares of the @xmath288 s to a slightly higher power ) , but a direct argument is actually easier . first we note that it suffices to show convergence in expectation . let @xmath167 be such that the statements in proposition [ prop : meyers estimate ] hold . by hlder s and chebyshev s inequalities we have @xmath289 since @xmath290 is bounded , jensen s inequality yields @xmath291^{{\mathchoice{\myffrac{p}{2 } in \scriptstyle } { \myffrac{p}{2 } in \scriptstyle } { \myffrac{p}{2 } in \scriptscriptstyle } { \myffrac{p}{2 } in \scriptscriptstyle } } } \biggr)\le c{\mathbb e}\bigl(\bigl|\nabla_{i_k}(t\cdot\psi_\lambda)(\cdot , x_k)\bigr|^p\bigr).\ ] ] it follows that @xmath292 the right - hand side is bounded uniformly in @xmath72 . using this in , the claim follows . we can now finish the proof of our main results : proof of theorems [ thm1 ] and [ thm2 ] from propositions [ prop - approx - corrector - l2][prop-2.8 ] the distributional convergence in is a direct consequence of the martingale central limit theorem whose conditions ( [ e:2.4a][e:2.5a ] ) are established in propositions [ prop - mclt1 ] and [ prop - mclt2 ] . the limiting variance @xmath60 is given by the right - hand side of , in agreement with . it remains to prove that @xmath57 whenever @xmath58 and the law @xmath32 is non - degenerate . suppose on the contrary that @xmath293 . then for each @xmath161 we would have @xmath294 and thus @xmath295 @xmath32-a.s . denoting @xmath296 , ( [ eqn - hdef][e:2.27 ] ) imply that , for @xmath32-a.e . @xmath178 , @xmath297\bigl|\nabla_i(t\cdot\psi)(\omega,0)\bigr|^2\big|{\mathcal f}_{(0,i)}\bigr)=0,\ ] ] where @xmath298 . let @xmath299 $ ] be the set of @xmath178 where this holds . then @xmath300 and , since @xmath32 is non - degenerate , @xmath301 contains at least two points . the expectation in is independent of @xmath302 ; subtracting the expression for two ( generic ) choices of @xmath178 in @xmath301 then shows that the inner integral must vanish for all @xmath303 . but tells us that the prefactor in square brackets , and thus the conditional expectation , is non - negative ( in fact , it is bounded away from zero ) . hence , this can only happen when @xmath304 but then @xmath305 , which can not hold for @xmath58 when is in force . proof of corollary [ cor1.2 ] from propositions [ prop - approx - corrector - l2][prop-2.8 ] thanks to ( [ e:2.1q][e:2.3 ] ) and proposition [ prop - mclt1 ] , @xmath61 is a martingale whose increments , @xmath75 are square integrable . therefore , @xmath306 but the right - hand side is the expectation of the quantity on the left of . since the convergence in occurs in @xmath238 , the claim follows . the goal of this section is to give proofs of propositions [ prop - approx - corrector - l2 ] and [ prop : meyers estimate ] . the former is a simple consequence of the hilbert - space structure underlying the definition of a harmonic coordinate ; the latter ( to which this section owes its name ) is a consequence of the caldern - zygmund regularity theory for singular integral operators . recall our notation @xmath120 for the operator in . we begin by noting an explicit representation of the minimum of @xmath128 as a function of the ( dirichlet ) boundary condition : [ lemma-4.1 ] let @xmath8 be finite and fix an @xmath119 . then there is @xmath307 , depending on @xmath10 and @xmath102 , such that for any @xmath235 that obeys @xmath308 for @xmath173 , @xmath309 ^ 2.\ ] ] moreover , @xmath310 for all @xmath311 and @xmath312 for all @xmath130 . proof `` integrating '' by parts we obtain @xmath313h(y ) . \end{aligned}\ ] ] employing the fact that @xmath235 is @xmath120-harmonic , the first sum drops out . for the second sum we recall that @xmath314 , where @xmath315 is the discrete poisson kernel which can be defined e.g. by @xmath316 for @xmath317 denoting the first exit time from @xmath10 of the random walk in conductances @xmath102 . now set @xmath318 and note that @xmath319 . it follows that @xmath320h(y ) = \sum_{y , z\in\partial\lambda } k(y , z)\bigl[h(y)-h(z)\bigr]h(y).\ ] ] the representation using the random walk and its reversiblity now imply that @xmath321 is symmetric . symmetrizing the last sum then yields the result . we note that lemma [ lemma-4.1 ] holds even for vector valued functions ; just replace @xmath322 ^ 2 $ ] by the norm squared of @xmath323 . this applies to several derivations that are to follow ; a point that we will leave without further comment . we can now prove proposition [ prop - approx - corrector - l2 ] dealing with the convergence of @xmath324 to @xmath269 in @xmath163-norm , as @xmath325 fills up all of @xmath0 . proof of proposition [ prop - approx - corrector - l2 ] abbreviate @xmath326 . the bound implies @xmath327 let @xmath328 be the minimizer of @xmath329 since @xmath235 is the minimizer of the corresponding dirichlet energy with conductances @xmath164 and boundary condition @xmath330 , we get using @xmath331 writing the last sum coordinate - wise and applying lemma [ lemma-4.1 ] , we thus get @xmath332 where the kernel @xmath333 pertains to the homogeneous problem , i.e. , the simple random walk . note that these bounds hold for all configurations satisfying . by shift covariance and sublinearity of the corrector ( cf proposition [ prop - corrector](2,4 ) ) , for each @xmath334 there is @xmath335 such that @xmath336 using this and in yields @xmath337 but @xmath338 for each @xmath339 while @xmath340 is , by lemma [ lemma-4.1 ] , the dirichlet energy of the function @xmath341 for conductances all equal to @xmath342 . hence , the last sum in is bounded by @xmath343 . taking @xmath218 and @xmath344 finishes the proof . [ r : representation ] as alluded to in the introduction , the @xmath152-convergence @xmath345 permits us to prove the formula for @xmath30 . the argument is similar to ( albeit much easier than ) what we used in the proof of proposition [ prop - mclt1 ] . indeed , we trivially decompose @xmath346 the stationarity of the gradients of @xmath143 and the spatial ergodic theorem imply that for any ergodic law @xmath32 on conductances , @xmath32-a.s . and in @xmath238 , @xmath347 it follows from the construction of the harmonic coordinate that expression on the right coincides with the infimum in . ( there is no gradient on the right - hand side of because @xmath147 . ) it remains to control the difference on the extreme right of . using the quadratic nature of @xmath26 , the ellipticity assumption and cauchy - schwarz , @xmath348 by proposition [ prop - approx - corrector - l2 ] which holds for any shift - ergodic ( elliptic ) law on conductances the right - hand side tends to zero as @xmath325 increases to @xmath0 . since we know that @xmath349 is bounded and converges almost surely ( e.g. , by the subadditive ergodic theorem ) , it converges also in @xmath238 . we conclude that the limit value @xmath30 is given by . key to the proof of proposition [ prop : meyers estimate ] is the meyers estimate . the term owes its name to norman g. meyers @xcite who discovered a bound on @xmath155-continuity ( in the right - hand side ) of the solutions of poisson equation with second - order elliptic differential operators in divergence from , provided the associated coefficients are close to a constant . the technical ingredient underpinning this observation is the caldern - zygmund regularity theory for certain singular integral operators in @xmath38 . ( incidentally , as noted in @xcite , meyers argument is a generalization of earlier work of boyarskii , cf ( * ? ? ? * and 3 ) for systems of first - order pdes and a version of his result was also derived , though not published , by caldern himself ; cf @xcite ) . to ease the notation , we will write @xmath350 for the canonical norm in @xmath351 , @xmath352 throughout the rest of this section . this carries no harm as all of our estimates will be pointwise rather than under expectation . let us review the gist of meyers argument for functions on @xmath0 . our notation is inspired by that used in naddaf and spencer @xcite , who seem to be the first to recognize its significance for the present type of problems , that in gloria and otto @xcite . a general form of the second order difference operator @xmath353 in divergence form is @xmath354 where @xmath355 are @xmath43-dependent matrix coefficients , @xmath356 is a vector whose @xmath161-th component is @xmath162 and @xmath357 is its conjugate acting as @xmath358 . the above @xmath353 is explicitly given by @xmath359-a_{i , j}(x-{\hat{\text{\rm e}}}_j)\bigl[f(x+{\hat{\text{\rm e}}}_i-{\hat{\text{\rm e}}}_j)-f(x-{\hat{\text{\rm e}}}_j)\bigr]\bigr).\ ] ] now , if @xmath360 is close to the identity matrix , it makes sense to write @xmath361 where we noted that the standard lattice laplacian @xmath362 corresponds to @xmath363 . this formula can be used as a starting point of perturbative arguments . consider a finite set @xmath8 and let @xmath364 . let @xmath14 be a solution to the poisson equation @xmath365 with @xmath366 on @xmath18 . employing , we can rewrite this as @xmath367.\ ] ] the function on the right has vanishing total sum over @xmath10 and hence it lies in the domain of the inverse @xmath368 of @xmath362 with zero boundary conditions . taking this inverse followed by one more gradient , and denoting @xmath369 this equation translates to @xmath370.\ ] ] a first noteworthy point is that this is now an autonomous equation for @xmath160 . a second point is that , if @xmath371 is the norm of @xmath372 as a map ( on vector valued functions ) @xmath373 and @xmath374 is the least a.s . upper bound on the coefficients of @xmath375 , uniform in @xmath43 , we get @xmath376 assuming @xmath377 this yields @xmath378 furthermore , the condition @xmath377 ensures the very existence of a unique solution @xmath160 to via a contraction argument ; then implies the continuity of @xmath379 in @xmath351 . the aforementioned general facts are relevant for us because @xmath120 is of the form . indeed , set @xmath380 and note that reduces to . the finite - volume corrector @xmath381 then solves the poisson equation @xmath382 thanks to , this @xmath40 is bounded uniformly so , in order to have for all finite boxes , our main concern is the following claim : [ lpbound - k ] for each @xmath170 , the operator @xmath383 is bounded in @xmath384 , uniformly in @xmath104 . proof of proposition [ prop : meyers estimate ] from theorem [ lpbound - k ] let @xmath385 . since ( in our setting ) @xmath386 , we may choose @xmath144 close enough to one so that @xmath387 . from the above derivation it follows @xmath388 we claim that this implies @xmath389 indeed , pick @xmath390 and note that , for any @xmath391 , @xmath392 taking expectations and dividing by @xmath393 , the left hand side becomes @xmath394 , while the second sum on the right can be bounded by @xmath395 , which tends to zero as @xmath218 by proposition [ prop - approx - corrector - l2 ] . applying chebyshev s inequality to the first sum on the right and taking @xmath218 followed by @xmath286 yields @xmath396 multiplying by @xmath397 and integrating over @xmath390 then proves . returning to the claims in proposition [ prop : meyers estimate ] , inequality is a restatement of . since ( [ e:4.23][e:4.24 ] ) imply the uniform boundedness of @xmath398 , for each @xmath399 , lemma [ lemma : p_p ] then shows @xmath400 , as @xmath218 for all @xmath399 . this proves as well . in the proof of theorem [ lpbound - k ] we will follow the classical argument by and large due to marcinkiewicz that is spelled out in chapter 2 ( specifically , proof of theorem 1 in section 2.2 ) of stein s book @xcite . the reasoning requires only straightforward adaptations due to discrete setting and finite volume , but we still prefer to give a full argument to keep the present paper self - contained . a key idea is the use of interpolation between the strong @xmath401-type estimate ( lemma [ l2bound - k ] ) and the weak @xmath402-type estimate for @xmath403 ( lemma [ weak-1 - 1-bound - k ] ) . both of these of course need to hold uniformly in @xmath104 . [ l2bound - k ] for any finite @xmath8 , the @xmath404-norm of @xmath405 satisfies @xmath406 . let @xmath407 be a hilbert space and @xmath408 a positive self - adjoint , bounded and invertible operator . then for all @xmath409 , @xmath410 we will apply this to @xmath407 given by the space ( of @xmath121-valued functions ) @xmath404 , @xmath411 and @xmath412 for some @xmath413 with zero boundary conditions outside @xmath10 . then @xmath414 where we used that @xmath357 is the adjoint of @xmath415 in the space of @xmath38-valued functions @xmath404 and where the various inner products have to be interpreted either for @xmath121-valued or @xmath38-valued functions accordingly . taking @xmath286 , the left - hand side becomes @xmath416 . the claim follows . the second ingredient turns out to be technically more involved . [ weak-1 - 1-bound - k ] @xmath383 is of weak - type ( 1 - 1 ) , uniformly in @xmath417 . that is , there exists @xmath418 such that , for all @xmath417 , @xmath419 and @xmath390 , @xmath420 deferring the proof of this lemma to the next subsection , we now show how this enters into the proof of theorem [ lpbound - k ] . we follow the proof in stein ( * ? ? ? * theorem 5 , page 21 ) . we begin with the case @xmath421 . let @xmath422 and pick @xmath390 . let @xmath423 and @xmath424 . then @xmath425 lemmas [ l2bound - k ] and [ weak-1 - 1-bound - k ] then yield @xmath426 with @xmath418 and @xmath427 independent of @xmath72 . multiplying by @xmath397 and integrating , we infer @xmath428 proving the assertion in the case @xmath421 . for @xmath429 , the facts that @xmath430 is obviously symmetric and that the norm admits the representation @xmath431 for @xmath142 equal to the index dual to @xmath268 , imply that @xmath432 . hence @xmath433 for all @xmath170 . it remains to prove lemma [ weak-1 - 1-bound - k ] . the strategy is to represent the operator using a singular kernel that has a `` nearly @xmath402-integrable '' decay . let @xmath434 be the green function ( i.e. , inverse ) of the laplacian @xmath362 on @xmath10 with zero boundary condition on @xmath18 . the operator @xmath372 admits the representation @xmath435=\sum_{y\in\lambda}\sum_{j=1}^d \bigl[\nabla_i^{(1)}\nabla_j^{(2)}g_\lambda(x , y)\bigr]f_j(y),\ ] ] where the superscripts on the @xmath415 s indicate which of the two variables the operator is acting on . since both @xmath436 and @xmath14 vanish outside @xmath10 , we have @xmath437=&\nabla_i\bigl(\,\sum_{y\in\lambda}g_\lambda(\cdot , y)\bigl(\nabla^\star\cdot f\bigr)(y)\big)(x)\\ = & \sum_{y\in{\mathbb z}^d}\bigl(\bigl(g_\lambda(x+{\hat{\text{\rm e}}}_i , y)-g_\lambda(x , y)\bigr)\sum_{j=1}^d[f_j(y-{\hat{\text{\rm e}}}_j)-f_j(y)]\bigr)\\ = & \sum_{j=1}^d\sum_{y\in{\mathbb z}^d}\bigl(g_\lambda(x+{\hat{\text{\rm e}}}_i , y+{\hat{\text{\rm e}}}_j)-g_\lambda(x , y+{\hat{\text{\rm e}}}_j)\bigr)f_j(y)\\ & \qquad\qquad\quad-\sum_{j=1}^d\sum_{y\in{\mathbb z}^d}\bigl(g_\lambda(x+{\hat{\text{\rm e}}}_i , y)-g_\lambda(x , y)\bigr)f_j(y ) . \end{aligned}\ ] ] this is exactly the claimed expression . crucial for the proof of the weak - type ( 1,1)-estimate in lemma [ weak-1 - 1-bound - k ] is an integrable decay estimate on the gradient of the kernel of the operator @xmath438 : [ green - decay - box ] there exists @xmath439 independent of @xmath72 such that @xmath440 for all @xmath441 and @xmath442 . although is certainly not unexpected , and perhaps even well - known , we could not find an exact reference and therefore provide an independent proof in section [ sec - greens - function - decay ] . with this estimate in hand , we can now turn to the proof of lemma [ weak-1 - 1-bound - k ] . to ease the notation , we will write @xmath325 ( note that all bounds will be uniform in @xmath72 ) and , resorting to components , write @xmath438 for the scalar - to - scalar operator with kernel @xmath443 for some fixed @xmath444 . for the most part , we adapt the arguments in stein @xcite . given a function @xmath445 , regard it as extended by zero outside @xmath10 . pick @xmath390 and consider a partition of @xmath0 into cubes of side @xmath446 , where @xmath447 is chosen so large that @xmath448 . naturally , each cube in the partition further divides into @xmath449 equal - sized sub - cubes of side @xmath450 , which subdivide further into sub - cubes of side @xmath451 , etc . we will now designate these to be either _ good cubes _ or _ bad cubes _ according to the following recipe . all cubes of side @xmath446 are _ ex definitio _ good . with @xmath452 being one of these sub - cubes of side @xmath450 , we call @xmath452 good if @xmath453 and bad otherwise . for each good cube , we repeat the process of partitioning it into @xmath449 equal - size sub - cubes and designating each of them to be either good or bad depending on whether holds or not , respectively . the bad cubes are not subdivided further . iterating this process , we obtain a finite set @xmath454 of bad cubes which covers the ( bounded ) region @xmath455 . we define @xmath456 , the good region , and note that @xmath457 and @xmath458 where the last inequality is due to the fact that the parent cube of a bad cube is good . next we define the `` good '' function @xmath459 the `` bad '' function , defined by @xmath460 , then satisfies @xmath461 since @xmath462 , as soon as @xmath463 the desired bound will hold . we will now show these bounds in separate arguments . considering @xmath40 first , we note that @xmath464 is bounded by a constant times @xmath465 . indeed , for @xmath466 let @xmath467 denote the bad cube containing @xmath468 . then @xmath469 by using on @xmath470 and on @xmath471 . by chebychev s inequality and lemma [ l2bound - k ] , @xmath472 note that this yields an estimate that is uniform in @xmath325 because @xmath473 by lemma [ l2bound - k ] . let us turn to the estimate in concerning @xmath203 . let @xmath474 be an enumeration of the bad cubes and let @xmath475 be the restriction of @xmath203 onto @xmath476 . abusing the notation to the point where we write @xmath477 for the kernel governing @xmath438 , from we then have @xmath478b(y),\ ] ] where @xmath479 is the center of @xmath476 ( remember that all cubes are odd - sized ) . let @xmath480 denote the cube centered at @xmath479 but of three - times the size i.e. , @xmath480 is the union of @xmath476 with the adjacent @xmath481 cubes of the same side . the bound now proceeds depending on whether @xmath482 or not . for @xmath483 , the distance between @xmath468 and any @xmath484 is proportional to the distance between @xmath468 and @xmath479 . proposition [ green - decay - box ] thus implies @xmath485 moreover , thanks to , @xmath486 using these in yields @xmath487 summing over all @xmath483 and taking into account that @xmath488 for @xmath482 , we conclude @xmath489 for some constant @xmath490 . setting @xmath491 and summing over @xmath71 , we obtain @xmath492 which by an application of markov s inequality yields @xmath493 i.e. , a bound of the desired form . to finish the proof , we still need to take care of @xmath494 . here we get ( and this is the only step where we are forced to settle on _ weak_-type estimates ) , @xmath495 the bound then follows by combining , and . in order to finish the proof of theorem [ lpbound - k ] , we still need to establish the decay estimate in proposition [ green - decay - box ] . this will be done by invoking a corresponding bound in the full lattice and reducing it onto a box by reflection arguments . ( this is the sole reason why we restrict to rectangular boxes ; more general domains require considerably more sophisticated methods . ) for @xmath334 , let @xmath496 denote the green function associated with the discrete laplacian @xmath362 on @xmath0 with killing rate @xmath334 , i.e. , @xmath497 is the kernel of the bounded operator @xmath498 on @xmath499 . this function admits the probabilistic representation @xmath500 where @xmath501 is the simple random walk and @xmath502 is the law of @xmath501 started at @xmath43 . this function depends only on the difference of its arguments , so we will interchangeably write @xmath503 . we now claim : [ green - eps - decay - lattice ] there exists @xmath504 such that , for all @xmath334 , all @xmath505 and all @xmath506 , @xmath507 this is a mere extension ( by adding one more gradient ) of the estimates from in lawler ( * ? ? ? * theorem 1.5.5 ) . ( strictly speaking , this theorem is only for the transient dimensions but , thanks to @xmath334 , the same proofs would apply here . ) the main idea is to use translation invariance of the simple random walk to write @xmath508 as a fourier integral and then control the gradients thereof under the integral sign . we leave the details as an exercise to the reader . we now state and prove a stronger form of proposition [ green - decay - box ] . [ green - eps - decay - box ] there exists @xmath439 such that , for all @xmath417 , @xmath334 and arbitrary @xmath505 , @xmath509 for all @xmath183 and all @xmath442 . here , the superscripts on the operators indicate the variable the operator is acting on . throughout , we fix @xmath510 and denote @xmath325 . the proof is based on the reflection principle for the simple random walk on @xmath0 . let @xmath511 denote the @xmath161-th component of @xmath501 and let @xmath512 for @xmath513 with components @xmath514 , and integer - valued indices @xmath515 , put @xmath516 our first claim is that , for @xmath517 , @xmath518 first we note that , the restriction on the length of @xmath501 makes this effectively a finite sum so we only have to exhibit appropriate cancellations due to the alternating sign . let @xmath519 , for @xmath520 and @xmath515 , denote the set of paths of length @xmath71 starting at @xmath43 , ending in @xmath521 and visiting @xmath522 exactly @xmath523 times . let @xmath524 if the path @xmath525 ends in @xmath526 and @xmath527 otherwise . the cancellation will arise from the fact that , @xmath528 to see this consider a map of @xmath529 , @xmath530 , onto itself by taking a path and reflecting the part after the last visit to @xmath522 . this is a bijection from @xmath529 onto itself which changes the sign of @xmath531 . since @xmath519 is finite , the sum must vanish . as all paths in @xmath529 have the same probability , we are permitted to multiply by the probability of each respective path and get for all @xmath530 that @xmath532}=p ) & = \sum _ { p \in a^k_m,\,n\in{\mathbb z } } ( -1)^ { s(p ) } p^x\bigl(x_{[k]}=p , x_k = r^i_n(y)\bigr)\\ & = \sum _ { p \in a^k_m,\,n\in{\mathbb z } } ( -1)^n p^x\bigl(x_{[k]}=p , x_k = r^i_n(y)\bigr)\\ & = \sum_{n\in{\mathbb z } } ( -1)^n p^x\bigl(x_{[k]}\in a^k_m , x_k = r^i_n(y)\bigr ) , \end{aligned}\ ] ] where @xmath533}$ ] denotes the path of the random walk up to time @xmath71 . we now verify by @xmath534}\in a^k_0\bigr)\\ & = \sum_{n\in{\mathbb z } } ( -1)^n p^x\bigl(x_{[k]}\in a^k_0,x_k = r^i_n(y)\bigr)\\ & = \sum_{\,n\in{\mathbb z}}\,\sum_{m\in{\mathbb n } } ( -1)^n p^x\bigl(x_{[k]}\in a^k_m , x_k = r^i_n(y)\bigr)\\ & = \sum_{n\in{\mathbb z } } ( -1)^n p^x\bigl(x_k = r^i_n(y)\bigr ) , \end{aligned}\ ] ] where was used in the third equality . ( there are no convergence issues as all sums remain effectively finite because @xmath71 is fixed . ) this obviously holds regardless of any restriction of the other components of the walk , which means that we have in particular @xmath535 for each @xmath536 . we are now ready to establish the desired representation for the green function . the argument proceeds by induction on dimension . abusing our earlier notation , denote @xmath537 for any @xmath538 , the green function @xmath539 on @xmath540 with zero boundary condition is given by @xmath541 applying to every probability term , we obtain for each @xmath536 @xmath542 consecutive applications of this equality yield @xmath543 for all @xmath183 , where we abbreviate @xmath544 . from lemma [ green - eps - decay - lattice ] , we thus obtain @xmath545 for all @xmath183 . let @xmath183 and abbreviate @xmath546 . whenever @xmath547 , we have @xmath548 as reflection always increases the distance between points in @xmath10 . if @xmath549 , we may even estimate @xmath550 the latter is verified quickly using @xmath551 and the fact that @xmath552 is at least @xmath553 in this case . therefore , we obtain @xmath554 which is the desired estimate . we are now ready to complete the proof of theorem [ lpbound - k ] : although the @xmath344 limit of @xmath555 exists only in @xmath64 , for gradients we have @xmath556 in all @xmath52 . since the bound in lemma [ green - eps - decay - box ] holds uniformly in @xmath334 , we get the claim in all @xmath52 . in this section we will prove propositions [ prop - perturbed - corrector ] and [ prop-2.8 ] . abandoning our earlier notation , let @xmath557 denote the green function in @xmath10 with dirichlet boundary condition for conductance configuration @xmath102 . ( thus , the simple - random walk green function from section [ sec - meyers ] corresponds to @xmath558 . ) the green function is the fundamental solution to the poisson equation , i.e. , @xmath559 \,\,\,g_\lambda(x , z,\omega)=0,\qquad&\text{if } z\in\partial\lambda , \end{cases}\ ] ] where @xmath560 is the kronecker delta . note that @xmath436 is defined for all @xmath119 . the solution to is naturally symmetric , @xmath561 and so we can extend it to a function on @xmath562 by setting @xmath563 whenever @xmath130 . here is a generalized form of the representation ( we thank a careful referee who pointed out that this result has appeared in a very similar form in @xcite , lemma 2.4 ) : [ lemma-5.1 ] for a finite @xmath8 let @xmath183 be nearest neighbors . for any @xmath185 such that @xmath564 except at @xmath186 , and any @xmath565 , @xmath566\bigl[\psi_\lambda(\omega , y)-\psi_\lambda(\omega , x)\bigr ] . \quad\end{gathered}\ ] ] proof suppose @xmath567 are such that @xmath212 equals @xmath102 except at the edge @xmath186 , where @xmath568 . define the function @xmath569 by @xmath570\bigl[\psi_\lambda(\omega , y)-\psi_\lambda(\omega , x)\bigr].\ ] ] we claim that @xmath571 since for @xmath572 we have @xmath573 , this will imply @xmath574 thanks to the uniqueness of the solution of the dirichlet problem . in order to show , we first use ( [ e:5.2][e:5.3 ] ) to get @xmath575\bigl[\psi_\lambda(\omega , y)-\psi_\lambda(\omega , x)\bigr].\ ] ] to deal with the term @xmath576 , we think of of @xmath577 as a matrix of dimension @xmath578 . for its coefficients @xmath579 we obtain @xmath580\bigl[\delta_y(z')-\delta_x(z')\bigr].\ ] ] using that @xmath581 for @xmath582 , we now readily confirm . proof of proposition [ prop - perturbed - corrector ] set @xmath583 and denote @xmath584 . lemma [ lemma-5.1 ] shows @xmath585\nabla_i\psi_\lambda(\omega , x),\ ] ] where the superindices on @xmath415 indicate which variable is the operator acting on . to prove the claim we need to show @xmath586^{-1}=\inf\bigl\{q_{\lambda}(f)\colon f(y)-f(x)=1,\,0\le f\le 1,\,f_{\partial\lambda}=0\bigr\},\ ] ] where the conductances in @xmath26 correspond to @xmath102 . for this , let @xmath14 be the minimizer of the right - hand side . the method of largrange multipliers shows @xmath587.\ ] ] thanks to , this is solved by @xmath588=\alpha\nabla_i^{(1)}g_\lambda(x , z;\omega)\ ] ] which in light of the constraint @xmath589 gives @xmath590^{-1}$ ] . since also @xmath591 , gives @xmath592 and so holds . the correspondence then follows from ( [ e:5.8d][e:5.9 ] ) . it remains to prove the equalities in . the first equality follows form by exchanging the roles of @xmath102 and @xmath212 . for the second equality , we note that relates the said factor to the ratio of double gradients of the green function at @xmath102 and @xmath212 which we abbreviate as @xmath593 . finally , it remains to establish the limit , including all of its stated properties : proof of proposition [ prop-2.8 ] thanks to ellipticity restriction , we have a bound on this quantity in terms of the lattice laplacian . keeping in mind the representation in , we have in fact that , for some @xmath594 related to the double gradient of the green function for all conductances equal to one , @xmath595 uniformly in @xmath10 . moreover , again by , @xmath596 is non - decreasing in @xmath10 and so the limit exists . the formula and the claimed stationarity then follow . this research was partially supported by the nsf grant dms-1106850 , nsa grant h98230 - 11 - 1 - 0171 and the gar project p201/11/1558 . two of the authors ( m.s . and t.w . ) gratefully acknowledge financial support from the esf grant `` random geometry of large interacting systems and statistical physics '' ( rglis ) that made their stay at ucla , and the work presented in this note , possible . the authors also wish to thank three anonymous referees for their valuable comments and suggestions on various aspects of this work .

given a resistor network on @xmath0 with nearest - neighbor conductances , the effective conductance in a finite set with a given boundary condition is the the minimum of the dirichlet energy over functions with the prescribed boundary values . for shift - ergodic conductances , linear ( dirichlet ) boundary conditions and square boxes , the effective conductance scaled by the volume of the box converges to a deterministic limit as the box - size tends to infinity . here we prove that , for i.i.d . conductances with a small ellipticity contrast , also a ( non - degenerate ) central limit theorem holds . the proof is based on the corrector method and the martingale central limit theorem ; a key integrability condition is furnished by the meyers estimate . more general domains , boundary conditions and ellipticity contrasts will be addressed in a subsequent paper . _ @xmath1department of mathematics , ucla , los angeles , california , u.s.a . _ _ @xmath2school of economics , university of south bohemia , esk budjovice , czech republic _ _ @xmath3institut fr mathematik , technische universitt berlin , berlin , germany _ _ @xmath4weierstra - institut fr angewandte analysis und stochastik , berlin , germany _ 0.3 cm

the widom rowlinson ( wr ) model , introduced in 1970 @xcite as an ingeniously simple model for the study of phase transitions in continuum fluids ( for an overview see ref . @xcite ) , continues to be , like its authors , a rich source of insights and analytical results in many ( sometimes quite unexpected ) areas @xcite of statistical mechanics . in this paper , dedicated with great pleasure to ben widom on the occasion of his seventieth birthday , we continue our study of a variation of the original model from two to @xmath0 components on a lattice : hard core exclusion between particles of different species on nearest neighbor sites . this model was first considered by runnels and lebowitz @xcite who proved that when the number of components @xmath0 is larger than some minimum @xmath11 then the transition from the gas phase at small values of @xmath1 to the demixed phase at large values of @xmath1 does not take place directly . instead there is , at intermediate values of @xmath1 , @xmath12 , an ordered phase in which one of the sublattices ( even or odd ) is preferentially occupied , i.e. there is a crystalline ( antiferromagnetically ordered ) phase in which the average particle density on the even and odd sublattices , @xmath13 and @xmath14 are unequal . the average density , @xmath15 , of species @xmath16 , on each sublattice is the same for each @xmath17 , with @xmath18 and @xmath19 . the nature of the symmetry breaking is thus very different from that in the demixed phase at @xmath20 where @xmath21 but there exists one species , say @xmath22 , for which @xmath23 . the origin of this crystalline phase is purely entropic . for @xmath1 fixed and @xmath0 large `` it pays '' for the system `` entropy wise '' to occupy just one sublattice without any constraint ; since there are no interactions between particles on the same sublattice there are @xmath0 independent choices at each site if we keep one of the sublattices empty . this more than compensates , at some @xmath24 , for the `` loss '' of `` fugacity energy '' occasioned by keeping down the density in one of the sublattices . a natural question now arises , just how big does @xmath11 have to be to see this ordered phase for @xmath25 . it was shown in @xcite that on the square lattice @xmath26 ; a ridiculously large upper bound . on the other hand a direct computation on the bethe lattice @xcite with @xmath27neighbors gives @xmath28 ^ 2 $ ] , which would suggest @xmath29 for the square lattice , @xmath30 for the cubic and @xmath31 for the bcc lattice . now it can be shown , using fkg inequalities , that @xmath32 on any bipartite lattice @xcite , but beyond that we have no simple or convincing argument for finding @xmath11 . we therefore turned to monte carlo simulations . this gave on the square lattice @xmath33 @xcite which is only about twice as large as the bethe lattice prediction . this wetted our appetite to try the bcc lattice where @xmath34 . to our surprise we find here , using monte carlo simulations , that @xmath11 does indeed equal @xmath35 , on the bcc lattice . while we have no clue of how to find rigorously the actual value of @xmath11 or of @xmath3 , it was argued in @xcite that for a given @xmath1 and @xmath0 large enough , the typical occupancy pattern on the lattice ( ignoring the label @xmath17 of the particles ) should be like that of a one component lattice gas with nearest neighbor hard core exclusion . for the latter system dobrushin @xcite proved the existence of a crystalline state . this implies that @xmath3 should behave for large @xmath0 as @xmath36 , where @xmath37 is the critical fugacity at which the one component hard cube gas ( occupation at a site @xmath38 excludes occupation at all eight neighboring sites ) crystalizes . the value of @xmath37 for the bcc lattice , obtained by gaunt @xcite using series expansion methods , is @xmath39 . using mc we obtained @xmath40 which is roughly consistent with gaunt s value . our result also agrees well with the values of @xmath41 for large @xmath0 being approximately @xmath42 for @xmath43 and @xmath44 . this provides solid evidence for the existence of a reentry phase transition in the temperature magnetic field plane phase diagram of an ising spin system with nearest - neighbor anti - ferromagnetic interactions on the bcc lattice @xcite . we also find , as in @xcite , that for large @xmath0 , @xmath4 for the crystal demixed transition can be computed via pirogov sinai theory @xcite yielding , @xmath45 which matches up smoothly with our mc results , see fig . it is easy to show that there is a demixing transition for @xmath46 @xcite , the existence of sharp interfaces between coexisting phases , for @xmath6 , on the cubic lattice at large fugacity @xmath1 was proven in @xcite . we next present results of our simulations and refer the reader to @xcite for a more detailed description of the model and additional references . our mc simulations were carried out on a bcc lattice of size @xmath47 with periodic boundary conditions . on an initially empty lattice we deposit particles chosen at random from the @xmath0 components at fugacity @xmath1 respecting the exclusion of different species occupying neighboring sites . we then sequentially update the lattice using a checkerboard algorithm resulting in a good vectorization . an update of a lattice site @xmath48 on one of the two simple cubic sublattices @xmath49 , making up the bcc lattice , which is occupied by a particle of type @xmath17 ( @xmath50 indicating an empty site ) is done as follows : we randomly choose a new trial particle of type @xmath51 , where @xmath51 can have any integer value between @xmath52 and @xmath0 with equal probability . @xmath53 refers to an attempted removal of a particle @xmath54 from the lattice site , which is successful , if a number @xmath55 randomly chosen with equal probability between @xmath52 and @xmath56 is smaller than the inverse fugacity @xmath57 . when this occurs @xmath17 gets the value @xmath52 , otherwise it remains unchanged , @xmath58 refers to an attempted deposition of a particle of type @xmath51 . if @xmath50 then it is successful if each of the four nearest neighbor sites is either empty or occupied by a particle of the same type ( @xmath51 ) and @xmath59 . in this case @xmath17 gets the value @xmath51 , otherwise it remains unchanged . a direct replacement attempt of a particle @xmath60 surrounded by eight empty nearest neighbor sites is always successful . typically in a simulation run after an equilibration of @xmath61 monte carlo steps ( mcs ) we update the lattice @xmath62 times , the configuration of every tenth step is taken for the evaluation of the averages . a typical run with @xmath62 mcs took about 1 cpu hours on a cray ymp . let @xmath63 denote the occupancy of a site , @xmath64 if the site @xmath48 is empty and @xmath65 otherwise . as observables we took histograms @xmath66 of the order parameter @xmath67 for the crystal structure and @xmath68 of the order parameter @xmath69 for the demixed phase in subsystems of size @xmath70 , @xmath71(-1)^{2*s-1 } \label{pc}\ ] ] and @xmath72 where @xmath73 denotes the number of particles of type @xmath17 in a subsystem of size @xmath74 and @xmath75 is the average overall density . for a given @xmath0 the transition activity @xmath76 is found by finite size scaling techniques @xcite . in particular the @xmath77 moments of the order parameter distribution @xmath66 , @xmath78 can be evaluated in subsystems of size @xmath74 , and from them the fourth order cumulant @xcite @xmath79 , in a one phase region far away from a critical point the subsystem size typically can be chosen larger than the correlation length @xmath81 , @xmath82 and the order parameter distribution is to a good approximation a gaussian centered around @xmath52 , resulting in @xmath83 for @xmath84 . in the two phase coexistence region far away from a critical point we can again assume @xmath85 and the order parameter distribution is bimodal resulting in @xmath86 for @xmath84 . near the critical point however we have @xmath87 , and using scaling arguments @xcite the cumulant is a function of @xmath88 , resulting for @xmath89 in the same value of @xmath90 for all different @xmath91 . this method allows the efficient determination of critical points by analyzing the cumulants for different values of @xmath1 on different length scales @xmath91 . applied to our model we should see , for low values of @xmath1 , when the system is in the disordered one phase region , @xmath92 for @xmath93 . increasing @xmath1 , we obtain , for large enough @xmath0 , a crystal phase with @xmath94 for @xmath95 . near @xmath76 we expect @xmath96 for @xmath97 . this method for locating the transition fugacities was used in our previous studies @xcite on the square lattice as well . for @xmath98 we obtain in this way values of @xmath99 respectively , for @xmath100 , @xmath101 and for @xmath102 , @xmath103 . cumulants for the gas crystal transition of the hard diamond and the @xmath104component system are shown in figs . 1 and 2 , respectively . the transition points are presented in fig . 3 together with the asymptotic expression @xmath105 . this corresponds to the value of @xmath37 on the bcc lattice @xmath40 which we obtained using mc techniques . we also show in fig . 3 a purely empirical fit of the @xmath106corrections for @xmath76 , @xmath107 , to the mc data , with a value of @xmath108 . the minimum number of components required for the existence of the crystal phase on the bcc lattice is @xmath109 . in fig . 4 we show the approach of the critical fugacity @xmath110 to the limit of the hard diamond system . for @xmath6 we obtain a direct transition from the gas phase to the demixed phase at a transition fugacity of @xmath111 ; the transition is second order with ising exponents as on a simple cubic lattice studied by dickman et al . @xcite . for @xmath112 we observe a direct first order transition from the crystal to the demixed phase . this transition was analyzed by studying the order parameter distribution @xmath68 . in the simulations we find a hysteresis region around @xmath113 , going approximately between two values of @xmath1 , say @xmath114 and @xmath115 , when increasing and decreasing the fugacity . in cases of a small hysteresis region with extent of less than @xmath116 , @xmath113 was taken as the average of this region , @xmath117 . in cases of a larger hysteresis region we located the transition fugacity @xmath113 by finding the relative stability of one of the two phases during a simulation starting from configurations with both phases present in parallel slices extending over the length of the simulation box , as described in @xcite . the resulting phase transition values of @xmath113 are shown in fig . 1 . we note that with increasing number of components , the transition fugacities approach the exact asymptotic line @xmath118 . we thank a. mazel for very helpful discussions . p.n . acknowledges support from the deutsche forschungsgemeinschaft ( heisenberg foundation ) , the computations were carried out at the cray - ymp of the rhrk kaiserslautern and the cray - t90 of the hlrz jlich . j.l.l . was supported by nsf grant dmr 95 - 23266 and dimacs and its supporting agencies , the nsf under contract stc-91 - 19999 and the n.j . commission on science and technology .

we use monte carlo techniques and analytical methods to study the phase diagram of the @xmath0component widom rowlinson model on the bcc lattice : there are @xmath0 species all with the same fugacity @xmath1 and a nearest neighbor hard core exclusion between unlike particles . simulations show that for @xmath2 there is a `` crystal phase '' for @xmath1 lying between @xmath3 and @xmath4 while for @xmath5 there are @xmath0 demixed phases each consisting mostly of one species . for @xmath6 there is a direct second order transition from the gas phase to the demixed phase while for @xmath2 the transition at @xmath4 appears to be first order putting it in the potts model universality class . for @xmath0 large , pirogov - sinai theory gives @xmath7 . in the crystal phase the particles preferentially occupy one of the sublattices , independent of species , i.e. spatial symmetry but not particle symmetry is broken . for @xmath8 this transition approaches that of the one component hard cube gas with fugacity @xmath9 . we find by direct simulations of such a system a transition at @xmath10 which is consistent with the simulation @xmath3 for large @xmath0 . this transition appears to be always of the ising type .